WO2022068931A1 - Procédé et appareil de reconnaissance d'empreintes digitales sans contact, terminal, et support d'informations - Google Patents

Procédé et appareil de reconnaissance d'empreintes digitales sans contact, terminal, et support d'informations Download PDF

Info

Publication number
WO2022068931A1
WO2022068931A1 PCT/CN2021/122240 CN2021122240W WO2022068931A1 WO 2022068931 A1 WO2022068931 A1 WO 2022068931A1 CN 2021122240 W CN2021122240 W CN 2021122240W WO 2022068931 A1 WO2022068931 A1 WO 2022068931A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
fingerprint
finger
fingerprint image
fingertip
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2021/122240
Other languages
English (en)
Chinese (zh)
Inventor
汤林鹏
邰骋
张青笛
王心安
刘勤
孙睿骅
金亦奇
王雪梅
胡伟
陈泽
仇卓
苏东泉
王晓航
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Moqi Technology (beijing) Co Ltd
Original Assignee
Moqi Technology (beijing) Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN202011069036.7A external-priority patent/CN112232163B/zh
Priority claimed from CN202011061125.7A external-priority patent/CN112232157B/zh
Priority claimed from CN202011056418.6A external-priority patent/CN112232152B/zh
Priority claimed from CN202011062601.7A external-priority patent/CN112232159B/zh
Priority claimed from CN202011056390.6A external-priority patent/CN112016525A/zh
Priority claimed from CN202011057702.5A external-priority patent/CN112232155B/zh
Application filed by Moqi Technology (beijing) Co Ltd filed Critical Moqi Technology (beijing) Co Ltd
Publication of WO2022068931A1 publication Critical patent/WO2022068931A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition

Definitions

  • the present application relates to the field of fingerprint identification, and in particular, to a non-contact fingerprint identification method, device, terminal and storage medium.
  • biometric authentication technology has been widely used in recent years, and related application scenarios such as face recognition, voiceprint recognition, etc.
  • Face-based identity authentication technology is developing rapidly. Whether it is 1:N face comparison or 1:1 identity verification, there are many landing scenarios, such as identity verification, gate pass, offline verification. payment etc.
  • biometric identification technology that accounts for more than 50% of the market share in biometric identification, fingerprints have extensive and in-depth applications in personal consumer electronics, security, financial banking and other fields.
  • traditional optical or capacitive contact fingerprint collection devices also have problems such as low collection quality, small collection area, sensitivity to skin dryness and wetness, and low collection consistency.
  • the present application is made to solve the above problems, and provides a non-contact fingerprint identification method, device, terminal and storage medium.
  • a non-contact fingerprint identification method which may include:
  • One or more fingers of the object to be recognized are photographed to obtain a finger image including the one or more fingers; the fingers of the one or more fingers in the finger image are recognized according to a finger recognition model tip positions to obtain one or more fingertip frames corresponding to the one or more fingers, the fingertip frames including the most distal ends of the one or more fingers to the one or more fingers
  • the fingertip area of the first phalangeal line of a finger; one or more corresponding fingertip positions of the one or more fingers are obtained from the finger image according to the one or more fingertip boxes a plurality of fingertip images; performing image processing on the one or more fingertip images to obtain a fingerprint image; and comparing the fingerprint image with a fingerprint library image to obtain a comparison result, comprising: locally Compare the fingerprint image with the fingerprint database image to obtain a comparison result; or send the fingerprint image to a server, so that the server compares the fingerprint image with the fingerprint database image to obtain a comparison result .
  • the non-contact fingerprint identification method may further include: acquiring incidental information of the one or more fingers, wherein the incidental information includes: concatenated finger information, finger position at least one of information, left and right hand information, mirror image information, and number of fingers information; and based on the additional information, determining a fingerprint image-finger position combination; wherein the fingerprint image is compared with the fingerprint database image to obtain Obtaining a comparison result includes: for each combination of the fingerprint image-finger position combination, comparing the fingerprint image and the fingerprint database image corresponding to the finger position corresponding to the fingerprint image under the combination.
  • the fingerprint database images are compared to obtain a comparison result; wherein, determining the fingerprint image-finger combination based on the incidental information includes: based on the incidental information, determining at least one fingerprint that satisfies the conditions of the incidental information Image-finger combination.
  • determining that at least one fingerprint image-finger position combination under the condition of the incidental information is satisfied may include: based on the incidental information, determining that the incidental information is satisfied All possible fingerprint image-finger combinations under information conditions.
  • the step of determining at least one combination of a fingerprint image and a finger position may include:
  • determining the order of the plurality of fingertip frames may include:
  • the order of the plurality of fingertip frames is determined in the direction from the start point to the end point.
  • the image processing on the one or more fingertip images may include at least one of preprocessing, normalization and image expansion, wherein the preprocessing Including removing the background, or adjusting the direction and removing the background: the normalization includes at least one of coarse normalization and fine normalization; the image expansion includes at least one of zoom-in, zoom-out, and image flattening; wherein, The direction adjustment includes: direction adjustment according to the direction of the first phalangeal line of the finger; direction adjustment according to the direction of a separate area; or direction adjustment according to the direction of the outer edge of at least one of the fingers including the knuckle ; The removing the background includes: removing the background according to the foreground area enclosed by the outline of the fingertip and the first phalanx line; or removing the background outside the finger area in a separate area; wherein, the first phalanx The line is determined by the knuckle line model, the fingertip contour line is formed by acquiring a plurality of boundary points
  • the coarse normalization may include:
  • Adjusting the size of the fingerprint image after removing the background according to the length of the first phalanx line or the area of the foreground area; and/or the fine normalization includes: detecting patterns of multiple sub-areas included in the fingerprint image line density; and adjusting the difference between the ridge line densities of the plurality of sub-regions to a preset density difference value; and/or calculating the overall frequency through the ridge line densities of the plurality of sub-regions, and adjusting the overall frequency to the overall preset value value.
  • detecting the ridge line density of the plurality of sub-regions included in the fingerprint image may include: determining the ridge line direction of the plurality of sub-regions;
  • the enlarging in the image expansion may include scaling up the normalized fingerprint image at least once to obtain an enlarged fingerprint image; the reducing in the image expansion includes scaling the normalized fingerprint image.
  • the unified fingerprint image is scaled down at least once to obtain a reduced fingerprint image.
  • the flattening in the image expansion includes: establishing a corresponding 3D fingerprint model according to the fingerprint image, and adjusting the fingerprint image according to the part corresponding to the fingerprint image in the 3D fingerprint model.
  • the fingerprint image is expanded to obtain a 2D fingerprint image corresponding to the fingerprint image.
  • obtaining the comparison result may include:
  • the weighted summation of the comparison scores of the fingerprint images corresponding to the plurality of finger positions under the combination is taken as the final comparison score under the combination;
  • the maximum among the corresponding final alignment scores determines the alignment result.
  • the comparison score of the fingerprint image corresponding to each finger position may be a fingerprint image corresponding to the finger position, an enlarged fingerprint image, a reduced fingerprint image, or a 2D fingerprint image.
  • the method may further include: a step of identifying a living body
  • the shooting of the one or more fingers of the object to be recognized includes: shooting one or more fingers of the object to be recognized under the condition of turning off the flash, obtaining a first finger image, and turning on the flash of the object to be recognized one or more fingers of the camera are taken to obtain a second finger image;
  • the living body identification step includes: performing Fourier transform on the first fingerprint image and the second fingerprint image to obtain a frequency domain signal; performing brightness analysis on each area of the first fingerprint image and the second fingerprint image to obtain The change value of the brightness of the corresponding areas of the first fingerprint image and the second fingerprint image when the flashlight is turned off and the flashlight is turned on; the stacked image obtained by stacking the first fingerprint image and the second fingerprint image, the frequency
  • the domain signal and the luminance value are input into a deep neural network, and a first judgment result representing whether the recognized object is a living body is obtained through the deep neural network; a recognition result is obtained according to at least one judgment result, and the at least one judgment result includes the first judgment result;
  • the comparing the fingerprint image with the fingerprint database image includes, when the recognition result shows that the recognized object is a living body, comparing the fingerprint image with the fingerprint database image; the first fingerprint The image is obtained by performing image processing on the image of the first finger; the second fingerprint image is obtained by performing image processing on the image of the second finger.
  • Embodiments of the present application also provide a non-contact fingerprint identification method, which may include:
  • Receive fingerprint data wherein the fingerprint data includes a fingerprint image and incidental information; based on the incidental information, determine all possible fingerprint-image-finger combinations that satisfy the condition of the incidental information.
  • the method may further include: if it is detected that the fingerprint image is collected without contact, performing coarse normalization and fine normalization on the fingerprint image; and/ Or fine normalizing the fingerprint image if it is detected that the fingerprint image was captured by contact.
  • the coarse normalization may include:
  • the size of the fingerprint image after removing the background is adjusted according to the length of the first knuckle line or the area of the foreground area, wherein the first knuckle line is determined by a knuckle line model, so
  • the fingertip outline is formed by acquiring a plurality of boundary points of the edge of the finger through an outline model and connecting the boundary points, and the foreground area is formed by the fingertip outline and the first finger. The area enclosed by nodal lines.
  • the fine normalization may include: detecting the ridge density of a plurality of sub-regions included in the fingerprint image; and
  • detecting the ridge line density of the plurality of sub-regions included in the fingerprint image may include: determining the ridge line direction of the plurality of sub-regions;
  • the embodiments of the present application also provide a device for non-contact fingerprint recognition, which may include:
  • an image acquisition module configured to: photograph one or more fingers of the object to be recognized to obtain a finger image including the one or more fingers;
  • an identification module configured to: identify fingertip positions of the one or more fingers in the finger image according to a finger identification model to obtain one or more fingers corresponding to the one or more fingers a fingertip frame comprising the fingertip area from the distal-most end of the one or more fingers to the first knuckle line of the one or more fingers; from the finger images according to the one or more fingertip frames obtain one or more fingertip images corresponding to fingertip positions of the one or more fingers;
  • a processing module configured to: perform image processing on the one or more fingertip images to obtain a fingerprint image
  • the comparison module is configured to: compare the fingerprint image with the fingerprint database image to obtain a comparison result, including:
  • the fingerprint image is sent to the server, so that the server compares the fingerprint image with the fingerprint database image to obtain a comparison result.
  • a terminal which may include:
  • At least one memory and at least one processor wherein, the at least one memory is used for storing program codes, and the at least one processor is used for calling the program code stored in the at least one memory to execute the above-mentioned method.
  • an embodiment of the present application further provides a storage medium, where the storage medium can be used to store program codes, and the program codes are used to execute the above method.
  • the technical solution of the present application can realize non-contact fingerprint recognition and improve user experience.
  • FIG. 1 is an exemplary system architecture diagram to which the present application can be applied;
  • FIG. 2 is a schematic diagram of a fingertip frame according to an embodiment of the present application.
  • FIG. 3 is a fingerprint image of a non-contact finger according to an embodiment of the present application.
  • FIG. 6 is an image captured by the same living finger under different light source colors according to an embodiment of the present application.
  • FIG. 7 is an image taken by the same living finger according to an embodiment of the present application under the conditions of different light source colors and polarizers;
  • FIG. 8 is an infrared image of a living finger according to an embodiment of the present application.
  • FIG. 9 is a schematic diagram of the application of a fingertip frame according to an embodiment of the present application.
  • FIG. 10 is a schematic diagram of an outline of an embodiment of the present application.
  • FIG. 11 is a schematic diagram of a non-contact fingerprint collection device according to an embodiment of the present disclosure.
  • FIG. 12 is a schematic diagram of another non-contact fingerprint collection device according to an embodiment of the present disclosure.
  • FIG. 13 is a schematic diagram of another non-contact fingerprint collection device according to an embodiment of the present disclosure.
  • FIG. 14 is a schematic diagram of a non-contact fingerprint collection device with an optical path adjustment device according to an embodiment of the present disclosure
  • FIG. 15 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
  • the term “including” and variations thereof are open-ended inclusions, ie, "including but not limited to”.
  • the term “based on” is “based at least in part on.”
  • the term “one embodiment” means “at least one embodiment”; the term “another embodiment” means “at least one additional embodiment”; the term “some embodiments” means “at least some embodiments”. Relevant definitions of other terms will be given in the description below.
  • the system structure 100 may include terminal devices 101 , 102 , 103 , and 104 , a network 105 and a server 106 .
  • the network 105 is a medium used to provide a communication link between the terminal devices 101 , 102 , 103 , 104 and the server 106 .
  • the electronic device (for example, the terminal device 101 , 102 , 103 or 104 shown in FIG. 1 ) on which the method of the present application operates can transmit various information through the network 105 .
  • the network 105 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others. It should be pointed out that the above wireless connection methods may include but are not limited to 3G/4G/5G connection, Wi-Fi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB connection, Local Area Network (“LAN”), Wide Area Network (“WAN”) ), the Internet (eg, the Internet), and peer-to-peer networks (eg, adhoc peer-to-peer networks), and other means of network connectivity now known or developed in the future.
  • the network 105 can communicate using any currently known or future developed network protocol, such as HTTP (HyperTextTransferProtocol), and can be interconnected with any form or medium of digital data communication (eg, a communication network).
  • HTTP HyperTextTransfer
  • the user can use the terminal devices 101, 102, 103, 104 to interact with the server 106 through the network 105 to receive or send messages and the like.
  • Various client applications may be installed on the terminal device 101, 102, 103 or 104, such as video live and playback applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platforms software, etc.
  • the terminal device 101, 102, 103 or 104 may be various electronic devices having a touch screen display and/or supporting web browsing, including but not limited to smartphones, tablet computers, e-book readers, MP3 (Motion Picture Expert Compressed Standard Audio) Layer 3) Player, MP4 (Motion Picture Expert Compression Standard Audio Layer 4) Player, Head Mounted Display Device, Notebook Computer, Digital Broadcast Receiver, PDA (Personal Digital Assistant), PMP (Portable Multimedia Player), Vehicle Terminals such as in-vehicle navigation terminals, etc., and mobile terminals such as digital TVs, desktop computers, and the like.
  • the terminal device 101, 102, 103 or 104 may also be a dedicated device dedicated to executing the method of the present application.
  • the server 106 may be a server that provides various services, such as a background server that provides support for pages displayed on the terminal device 101 , 102 , 103 or 104 or data transmitted.
  • terminal devices, networks and servers in FIG. 1 are merely illustrative. There can be any number of terminal devices, networks and servers according to implementation needs.
  • the terminal device can independently or cooperate with other electronic terminal devices to run applications in various operating systems, such as the Android system, to implement the implementation methods of the present application, and can also run applications in other operating systems, such as the iOS system, Windows system, HarmonyOS, etc.
  • the application of the system or the like realizes the method of the embodiment of the present application.
  • the non-contact fingerprint identification method of the embodiment of the present application may include the following steps.
  • the finger recognition step of the embodiment of the present application may include: acquiring a finger recognition model; judging whether the image includes the finger according to the finger recognition model; and when the image includes the finger, determining the finger The positions of the respective distal phalanx segments; wherein the orientation of the fingers includes any direction parallel to the image.
  • the step of obtaining the finger recognition model may include: obtaining a sample set, wherein the sample set may include fingers collected under different conditions and subjected to data augmentation processing (such as translation, rotation, noise addition, smoothing, etc.)
  • the different conditions may include different angles, different distances, different lighting, different sharpness and different collection devices; and the finger recognition model is obtained by training according to the finger images.
  • the step of judging whether the image includes the finger according to the finger recognition model may include: scanning the image by using a plurality of recognition frames with different areas, wherein, in each of the recognition frames scanned During the process, when the reliability of the finger contained in the area defined by the recognition frame is greater than the preset value, the recognition frame is retained in the area and continues to scan until the image is scanned, so that the Each of the reserved recognition frames includes at least a part of the finger; sort the reserved recognition frames according to the reliability, and obtain a sorting result; according to the sorting result, select the reserved recognition frames from the Get the fingertip frame so that one finger corresponds to only one fingertip frame.
  • the embodiments of the present application may provide a calling function based on the SDK form.
  • the upper-layer application calls this function, it can call up a collection page.
  • the presence or absence of a finger can be captured from each frame of image.
  • the human hand is placed at the set position and stabilized, the current image can be obtained, and through a series of models and calculations, the data containing the finger image is returned.
  • the model based on deep learning training can perform inference calculation for each frame of video stream in real time, and detect whether there are fingers and finger positions in the image.
  • a fingertip frame when entering the preview image, for each frame of image, through the model and calculation, determine whether A fingertip frame can be identified, wherein the identified fingertip frame refers to a retained and deduplicated fingertip frame, and the reliability of the retained fingertip frame is greater than a threshold; the fingertip frame is identified Indicates that the finger has been recognized. If the fingertip frame is not recognized for a period of time, the user can be interactively prompted that no finger is found in the image; if the finger is not recognized for a longer period of time, an error can be reported over time.
  • identifying a finger may include, but is not limited to, identifying and distinguishing different fingers; the fingertip frame used for identifying can accommodate the fingertip area from the fingertip to the first knuckle line, and may not accommodate the rest of the finger area; the finger It can be freely rotated in the image without affecting the correct recognition; the left-hand and right-hand fingers are automatically recognized according to the distinguishing characteristics of the fingers.
  • the supported identified finger combinations may include: thumbs of two hands, five fingers of one hand, four fingers of one hand, one finger of one hand, etc.
  • the finger recognition may include finger detection and finger tracking.
  • a preview image is obtained by a camera device, and the finger detection model inference calculation is performed on the first frame image, for example, the image is divided into several frames with a combination of length and width, and each frame has a reliability indicating the possibility of whether it contains a finger;
  • the above process obtains several fingertip frames, sorts all the corresponding fingertip frames in the image based on the reliability of the fingertip frames, and merges the fingertip frames through business logic (such as a 300x300 fingertip frame and a 310x310 fingertip frame).
  • Fingertip frame two fingertip frames contain the same finger), after the fingertip frame merging process is performed, one finger corresponds to only one fingertip frame, and the identified fingertip frame is returned.
  • finger tracking can be performed. For example, using the finger position of the previous frame image, that is, the position of the fingertip frame, using the finger recognition model to accurately locate the fingertip frame within the target position range, and determine whether it is a finger by the reliability of the fingertip frame, and returns the identified fingertip box. Finger tracking can further improve the recognition accuracy while reducing the time required for single-frame image recognition.
  • a fingerprint acquisition step in the case that the existence of the finger is detected in the finger identification step, focus and photograph the finger, and acquire respective fingerprint images of the fingers.
  • the fingerprint acquiring step may include: acquiring consecutive frames of images including the finger, detecting the position of the finger in the consecutive frames of images, and judging whether the variation range of the position is within a preset threshold range If the variation range of the position is within the preset threshold range, the respective focus states of the fingers are detected; if the fingers are in focus at the same time, a corresponding focal length is taken to obtain an image; if the fingers are not at the same time Focusing (for example, each finger is not on the same focal plane), the images with each finger as the focus point are separately captured, thereby obtaining multiple images.
  • each of the plurality of images includes all the fingers to be detected, since each of the plurality of images is in focus with at least one of the fingers, from the plurality of images
  • the image of the position of the fingertip frame corresponding to the finger in focus in the image is obtained from each image, so as to obtain the fingertip image corresponding to each finger.
  • the position of the finger can be displayed in the image in the form of a frame, as shown in Figure 2.
  • the finger needs to be stably detected from the preview image, and the finger needs to be in a proper position. For example, if the finger is too close to the camera device, there may be problems such as blurred focus and incomplete shooting of the finger. At this time, it can be judged whether the distance of the finger is too close according to the resolution of the camera device and the proportion of the finger in the image. , and prompts the user to move the finger further away when the finger is too close. On the other hand, if the finger is too far away from the camera, it may cause problems such as focusing on the background and unclear lines due to the limitation of resolution.
  • the ratio is used to judge whether the distance between the fingers is too far, and when the distance between the fingers is too far, the user is prompted to move the fingers closer.
  • the definition of the fingers is relatively higher when the fingers are closed. Because when the fingers are separated, the fingers that are far apart are prone to blurry shots due to the problem of misalignment of the focus plane. At this time, the distance between the fingers can be determined according to the position of the fingertip frame, and whether the fingers are closed is determined by logical operation, and when it is detected that the fingers are not closed, the user can be prompted to close the fingers. When the finger is in the proper position, the current image can be acquired.
  • the camera can be called by calling the camera interface to take a picture, or a frame image of the preview video stream can be directly used.
  • Both of these two forms of image acquisition can support active and automatic image acquisition. model.
  • the active acquisition can be that the operator manually clicks the button to acquire the relevant images;
  • the automatic acquisition can calculate whether the position change of the finger is within a certain threshold range according to the continuous multi-frame images, and judge whether the hand is relatively stable; if the acquisition device is If a mobile device (such as a mobile phone) has an acceleration gyroscope, you can also read the gyroscope information to determine whether the device is relatively stable. When it is judged that the device is stable, image acquisition can be automatically triggered.
  • the image transmission step of the embodiment of the present application may include: assigning a key to the fingerprint image, wherein the key is used to verify the key when receiving an access request for the fingerprint image, and when the fingerprint image is accessed When the key verification is passed, the access request is passed, and when the key verification fails, the access request is rejected. More specifically, the image transmission step of the embodiment of the present application may further include: before transmitting the encrypted fingerprint image to the comparison system, selecting a transmission interface, wherein the transmission interface includes fingerprint image storage interface, single-finger comparison interface, and multi-finger comparison interface.
  • the embodiments of the present application can perform encryption processing on the fingerprint image. For example, before calling the collection interface, a one-time key can be obtained from the background of the comparison system, and the key can have a certain validity period; when the collection interface is called, the key can be passed in. The key can be verified when shooting, and if the verification is passed, shooting can begin. After the shooting is completed, the captured data can also be encrypted using this key. More specifically, a one-time token (OTP) can be generated by using the secret key provided to the integrating party, and after the collection is completed, the encrypted fingerprint information can be generated by using the OTP, which is safe in transmission and prevents tampering.
  • OTP one-time token
  • the embodiment of the present application can encrypt the transmission and storage of the whole process data, separate the sensitive image data and the insensitive feature file data, control the access of the image data, and hash the feature file data to ensure the hash
  • the later features can still be compared with high precision, but no original information can be recovered from the hashed feature file, thus ensuring the security of the fingerprint image data.
  • a fingerprint comparison step comparing the fingerprint image with the fingerprint image in the comparison system to obtain a comparison result.
  • the fingerprint comparison step of the embodiment of the present application may include fingerprint authentication comparison and fingerprint search comparison.
  • the fingerprint matching capability can be provided in the form of an open API (application programming interface), wherein the API can include three interfaces: fingerprint image storage, one-to-one comparison of fingerprint images, and one-to-many comparison of fingerprint images.
  • the one-to-one comparison of fingerprint images may be 1:1 comparison between the collected fingerprint image and the fingerprint image of a known ID in the background fingerprint database to complete identity verification; the one-to-many comparison of fingerprint images may be a combination of The collected fingerprint image is 1:N compared with the fingerprint image in the background fingerprint database to check whether the fingerprint exists, etc.
  • the comparison system can be set up in a local or cloud server.
  • the embodiments of the present application may further include an image processing step of processing the fingertip image obtained in the fingerprint obtaining step, and the image processing step may include: obtaining each finger in the finger according to a knuckle line model. The two end points of the distal interphalangeal joint line of each finger, adjust the knuckle line to a preset direction according to the two end points; obtain a plurality of boundary points of the edge of each finger according to the finger contour model; connecting the two end points of each finger and the plurality of boundary points to form the outline of each finger; and extracting the outline from each of the plurality of images to obtain the Fingerprint image.
  • the embodiments of the present application may further include: connecting each fingertip frame to form a convex function; and classifying the fingers as left-handed according to the relationship between the convex function and the fingers (for example, the thumb is the farthest and the little finger is the shortest). Fingers or fingers of the right hand, and frame each fingertip as thumb, index finger, middle finger, ring finger, and little finger respectively.
  • the image obtained from the camera device can be used for finger recognition and image segmentation through the model to obtain the fingerprint image of each finger.
  • the finger recognition model can be used to obtain one or more fingertip frames, that is, the position of each finger, so as to generate one or more fingertip images.
  • Use the knuckle line model to obtain the positions of the two end points of the first knuckle line in the fingertip image, see A and B in Figure 2; use the finger contour model to obtain several boundary points of the finger edge.
  • the boundary points are connected with the knuckle line to form the outline of the finger. Based on the endpoints on both sides of the knuckle line and several points that form the outline of the finger, adjust the direction of the finger and remove the background of the finger through processing.
  • the orientation adjustment is performed by the endpoints on both sides of the knuckle line. Adjusting to a horizontal state can ensure that the fingertips of the fingers are facing upwards in the final output image. In this way, the hand can be rotated during shooting, and then the direction of the knuckle line can be adjusted to ensure the same direction of the output fingers; obtained through the finger contour line The foreground area where the finger is, and the background area outside the finger is removed.
  • Each fingertip frame is connected to form a convex function. Through the relationship between this function and the fingers (for example, the thumb is the farthest, the little finger is the shortest), it can be calculated and judged whether the photographed hand is the left hand or the right hand, and the corresponding value of each fingertip frame can be obtained.
  • the finger image can be scaled and output uniformly into a 500dpi image. For example, through statistics of a certain number of fingerprint images, the average value of the fingerprint area and the length of the knuckle line corresponding to different finger positions can be obtained. Adjust the image scale of each finger image; and calibrate and analyze several devices to obtain the relationship between the focal length read by the device and the actual object distance.
  • the finger image is captured by a camera, because the finger is a three-dimensional object, and the captured image is two-dimensional, which may lead to inconsistent ridge density in different areas of the fingerprint: the middle part of the fingerprint is sparse, and the surrounding ridges are sparse. Lines are denser.
  • the finger image can also be normalized.
  • a rough ridge density can be obtained.
  • the density is at or near a fixed value.
  • the depth information can also help us to expand the edge of the fingerprint, so that the center area of the fingerprint and the edge area have the same ridge density.
  • the image of the finger should be relatively clear. Therefore, quality judgment can be added in the form of image processing, including equalizing the image and filtering the image according to the frequency domain signal corresponding to the fingerprint lines. , Calculate the ridge signal in the figure, judge the overall image quality, etc.
  • the non-contact fingerprint identification method of the present application has been described as a whole above.
  • some people steal the user's fingerprint and imitate the finger with rubber, etc., to deceive the fingerprint recognition device, causing security risks. Therefore, in the non-contact fingerprint identification method of the present application, it is preferable to perform biometric identification before final fingerprint identification, and to perform fingerprint identification only when the identification object is a living body.
  • the living body identification method in the non-contact fingerprint identification method of the present application will be described with specific formulas.
  • the method of the present application includes the following steps.
  • the acquiring step acquiring a fingerprint image of at least a part of the fingerprint of at least one finger of the recognized object through a non-contact photographing method.
  • the finger when acquiring a fingerprint image of a finger, the finger is in contact with a fingerprint acquisition device (such as a fingerprint card punching machine), and the fingerprint image is acquired in a contact-type manner, and the relevant contact-type fingerprint image is binarized
  • a non-contact photographing method is used to obtain the fingerprint image, the finger is not in contact with the device for collecting the fingerprint image, and the collected fingerprint image may be a color fingerprint image.
  • At least one judgment result is obtained by judging whether the identified object is a living body in at least one way, and the identification result is obtained according to the at least one judgment result.
  • At least one method is used to determine whether the recognized object is a living body, and the recognized object is the owner of the fingerprint image of at least a part of the fingerprint of the at least one finger.
  • One determination method corresponds to one determination result.
  • multiple judgment methods can be used, and multiple judgment results can be obtained. Therefore, in some embodiments, multiple judgment results can be used to obtain identification results.
  • the recognition result is the final result of whether the recognized object is a living body obtained by the living body recognition step, that is, the judgment result is used as the basis for determining the recognition result.
  • the recognition result and the judgment result can be the same ;
  • the fingerprint identification step in the case where the living body identification result indicates that the identified object is a living body, fingerprint identification is performed on the fingerprint image.
  • whether to execute the fingerprint recognition result will be determined according to the recognition result. If the fingerprint recognition result shows that the recognized object is a living body, the case where the finger is a fake non-living finger is excluded. Performs fingerprinting, which increases security and reduces unnecessary computation.
  • the content of acquiring the fingerprint image has been specifically described in the above-mentioned part of the non-contact fingerprint identification method, and only the difference is described here.
  • the description in the above-mentioned part please refer to the description in the above-mentioned part.
  • the obtaining step includes: obtaining a fingertip image of at least one finger of the recognized object
  • the living body identification step includes: inputting the fingertip image into a trained neural network model, and obtaining the obtained image according to the output of the neural network model. the first result.
  • a finger image of a non-contact fingerprint is taken, and one or more fingertip positions are located through a finger tracking and detection algorithm.
  • the fingertip positions can be Including the fingerprint area, for the image including the fingerprint area, input it into the trained neural network model for processing, and the output result of the neural network model is used to represent whether the input image is the result of a living body, that is, the first result is obtained.
  • a non-contact fingerprint image of a living body and a 2D or 3D fingerprint image of a non-living body are collected in advance
  • x_i can be set to represent the ith fingerprint image
  • the established neural network model needs to be trained.
  • a classifier F of a neural network including a convolutional layer, a pooling layer and a fully connected layer is trained, and F is optimized by stochastic gradient descent or other optimization methods, so that F(x_i ) is associated with y_i, for example, the following formula can be used to train the classifier: max_ ⁇ F ⁇ 1/N* ⁇ sum_iy_i log F(x_i)+(1-y_i)(1-log F(x_i)).
  • F(x_i) means the probability that x_i is a non-contact fingerprint image of a living body. If F(x_i)>0.5, it means that x_i is likely to be a non-contact fingerprint image of a living body. If F(x_i ) ⁇ 0.5, it means that x_i is not a non-contact fingerprint image of a living body with high probability. In some embodiments, 0.5 is used as the judgment threshold by default to compare with F(x_i) to determine whether the fingerprint image is a living fingerprint image, and the false alarm rate and the false alarm rate of the system can also be adjusted by adjusting the judgment threshold, so that Adapt to the needs of the scene.
  • a non-living contactless fingerprint image is generated by at least one of data augmentation and an adversarial generative network (GAN), and a neural network model is trained by using the generated non-living contactless fingerprint image .
  • GAN adversarial generative network
  • a neural network model is trained by using the generated non-living contactless fingerprint image .
  • data augmentation and adversarial generation network are used to generate more non-living non-contact fingerprint images and reduce the over-fitting of the neural network. The accuracy of determining whether the fingerprint image is a living fingerprint image in practical applications is improved.
  • the acquiring step may include: acquiring images of the non-contact fingerprint of at least one finger of the recognized object under different lighting conditions.
  • the step of identifying the living body may further include: obtaining a third result representing whether the identified object is a living body according to images under different lighting conditions.
  • the at least one judgment result further includes: a third result.
  • the lighting conditions include at least one of brightness, light source color, and light source polarization state; images under different lighting conditions include: images obtained under different brightness, images obtained under different light source colors at least one of the images obtained under different polarization states of the light source.
  • Some non-living recognized objects may be finger images on printed cardboards or finger images displayed on electronic screens, so images with and without the flash can be taken separately to obtain images with different brightness, due to the cardboard.
  • the electronic screen has the characteristics of approximate plane reflection, so the accuracy of living body recognition can be improved by changing the brightness.
  • the physical characteristics of living and non-living bodies are different, so the display states under different light sources are different, and the accuracy of living body recognition can be improved by changing the color of the light source.
  • the light source can emit polarized light and non-polarized light, and it is determined whether it is a living body according to the difference in the reflected light of the object to be identified under the polarized light.
  • the brightness of the fingerprint image can be changed, the color of the light projected on the finger of the recognized object, and the polarization state of the light projected on the recognized object can be changed.
  • Fingerprint images taken under different brightness, fingerprint images taken under different light source colors, and fingerprint images under different light source polarization states can be obtained respectively. According to the three sub-judgment results, the third result is obtained, so that the accuracy of the judgment can be improved by synthesizing the sub-judgment results under various lighting conditions.
  • the reflectance of images obtained under different lighting conditions is calculated, and a third result indicating whether the recognized object is a living body is obtained according to the reflectance.
  • human skin has specific optical characteristics, it absorbs light in a specific wavelength range and generates light in a specific wavelength range.
  • the light reflected by non-human skin and the light reflected by human skin are not exactly the same. Therefore, the reflectivity can be compared by comparing the reflectance Determine if the identified object is alive. Since the reflectivity of some materials and the human body may partially overlap in certain wavelength ranges, but not in all wavelength ranges, therefore, optionally, images captured under different light source colors can be obtained separately, and different light sources can be calculated.
  • the reflectivity of the identified object under the color so as to determine whether the identified object is a living body according to the reflectivity under different light source colors.
  • the image under different light source color can be a video image under the light with continuously changing wavelength within a wavelength range, Thereby, the continuous change of reflectance with wavelength is obtained. At this time, the curve of reflectance with the wavelength of light is obtained. Compared with the use of reflectance point values, the use of this continuous reflectance change curve can better reflect the living body.
  • the fingerprint image is distinguished from the non-living fingerprint image.
  • images obtained under different lighting conditions are input into a deep neural network, and a third result representing whether the recognized object is a living body is obtained through the deep neural network.
  • a neural network model can be used for identification, wherein, taking the lighting conditions as the brightness as an example, the images taken by the finger of the living body under the condition of turning off the flash and turning on the flash are pre-collected, and the non- The images taken by the finger of the living body under the condition of turning off the flash and turning on the flash, wherein the image obtained from the image taken with the flash turned off is X_i ⁇ 1, and the image obtained from the image taken with the flash turned on Denoted as X_i ⁇ 2, i is a variable used to distinguish images of different objects, an image of an object can be denoted as (X_i ⁇ 1, X_i ⁇ 2), and (X_i ⁇ 1, X_i ⁇ 2) has a corresponding y_i, When the object is a living body
  • the method before inputting the fingerprint images obtained under different lighting conditions into the deep neural network, the method further includes: performing Fourier transform on the fingerprint images obtained under different lighting conditions to obtain frequency-domain signals, and converting the frequency-domain signals into The signal is also fed into the deep neural network, which improves the accuracy of the model's living body recognition.
  • the applicant of the present application has found that the frequency domain signals of images captured by living and non-living objects under different lighting conditions have a very high degree of discrimination, and by inputting the frequency domain signals into the deep neural network, the frequency domain signals for living and non-living objects can be significantly improved.
  • the deep neural network in this embodiment refers to the frequency domain signal to identify the living body and the non-living body, which greatly improves the accuracy of the identification compared with the related art.
  • fingertip images obtained under different lighting conditions are input into a support vector machine model, and a third result representing whether the recognized object is a living body is obtained through the support vector machine model.
  • inputting the fingertip images obtained under different lighting conditions into the deep neural network includes: performing brightness analysis on each area of the fingertip images obtained under different lighting conditions, calculating the The brightness change value under the condition, input the brightness change value into the deep neural network, or the support vector machine model, can also significantly improve the accuracy of living body recognition.
  • the obtaining step may include: obtaining an infrared image of the identified object, and the living body identifying step may further include: obtaining a fourth result indicating whether the identified object is a living body according to the infrared image.
  • the at least one judgment result may further include: a fourth result.
  • the skin on the surface of the human finger has a low reflectivity for infrared light, and the infrared light can partially penetrate the skin of the finger. It is an important difference between a living body and a non-living body, and these differences are reflected in the captured infrared images, so it can be determined whether the identified object is a living body according to the infrared image of the identified object.
  • obtaining the fourth result representing whether the recognized object is a living body according to the infrared image may include: obtaining optical properties of the recognized object under infrared light according to the infrared image, and according to the infrared image Optical properties under light yield the fourth result.
  • the optical properties of living bodies and non-living bodies under infrared light are different, which is manifested in the low reflectivity of the skin on the surface of human fingers to infrared light, and infrared light can penetrate part of the skin of the finger. Therefore, according to the captured infrared image
  • the reflectivity or transmittance of the identified object to infrared light can be determined, and whether the identified object is a living object is determined according to the reflectivity or transmittance, and a fourth result is obtained.
  • obtaining the fourth result indicating whether the recognized object is a living body according to the infrared image may include: extracting vein features from the infrared image, and performing a comparison between the vein features and pre-obtained living vein features The fourth result is obtained by comparison. Specifically, since infrared light can penetrate the surface skin of the human body, if the recognized object is a living body, the finger veins under the human skin can be photographed in the infrared image, and the vein features of the finger veins can be extracted from the infrared image.
  • the obtained vein features of the human body are compared, and if the degree of agreement between the two is higher than the preset value, it can be determined that the recognized object is a living body; if the degree of agreement between the two is lower than the preset value, it is determined that the recognized object is a non-living body , so as to obtain the fourth result.
  • a deep neural network can be used to identify infrared images.
  • the target object corresponding to the image can be determined according to the fingerprint image of the finger of the recognized object, and the pre-stored vein feature of the target object is used as the pre-obtained living vein feature. The fingerprint and vein features of the identification object are combined to judge, thereby improving the accuracy of judgment.
  • the fingerprint and vein features of employees are stored in the fingerprint punching machine at the same time, and the infrared image and fingerprint image are taken when punching the card. Only when the vein features and fingerprints in the fingerprint image correspond to the vein features and fingerprints stored by the same employee, the punch-in is considered successful.
  • the method in an embodiment of the present application is used in a mobile phone as an example for description below, and the method in the present embodiment may be implemented by an application in the mobile phone.
  • fingerprint images of multiple fingers of the object to be recognized are photographed, and multiple methods are used to perform determination based on the fingerprint images simultaneously, that is, there are multiple determination results, and the identification results are obtained by combining the multiple determination results.
  • the fingerprint image of the recognized object is first captured, and then three judgment results are obtained according to the color fingerprint image of the recognized object, the image of the recognized object under different brightness conditions, and the rPPG signal of the recognized object.
  • the identification results are obtained from the three judgment results, which will be described in detail below.
  • the camera device of the mobile phone locate the finger area of the object to be recognized and take an image of the finger area, and identify the fingertip area according to the captured image to obtain the fingertip image.
  • the fingertip image contains the fingerprint area, you can refer to the figure 3.
  • Input the obtained image into the deep neural network model, and the deep neural network model outputs the probability that the fingerprint image is a living fingerprint image.
  • the fingerprint image of at least one finger of the object to be identified is a color image.
  • the traditional contact fingerprint identification technology is a contact type total reflection imaging, and the collected image is a binarized image. Additional information It contains less information, and it is difficult to use fingerprint images for effective living identification; non-contact fingerprint images are imaged by RGB cameras, and contain greatly rich information, which makes it possible to detect living fingerprints by image-based deep learning methods.
  • the camera device of the mobile phone captures the finger image of the object to be recognized, and then turns on the flash to capture the finger image of the finger again, wherein the finger images of multiple fingers can be captured at the same time.
  • the image or fingerprint image obtained from the captured finger image is input into the deep neural network model, and the judgment result of whether the recognized object corresponding to the fingertip image or fingerprint image is a living body is output.
  • the non-living fingerprint image may be a fingerprint image printed on a cardboard or a fingerprint image displayed on an electronic screen. Since the cardboard image and the electronic screen image have the characteristics of plane or near plane reflection of light, the diffuse reflection of the light is generated. Reflection or specular reflection is visually distinct from the three-dimensional body of a living finger, especially since the rear camera device of the mobile phone is equipped with a flash, and when the flash is turned on, the difference is particularly obvious. Therefore, the different optical properties of the finger when the flash is turned on or off are analyzed, and two fingerprint images are stacked together and a deep neural network is used to determine the living body. In this way, the non-contact fingerprint image can be more accurately detected on the mobile phone without using additional lighting equipment.
  • the mobile phone captures the video stream of the finger of the recognized object, continuously tracks the position of the finger in the video stream, calculates the rPPG signal of one or more fingers, and obtains the judgment result according to the rPPG signal.
  • rPPG measures subtle brightness changes in the skin of an identified subject by analyzing reflected ambient light.
  • the subtle brightness changes of the skin are caused by the blood flow caused by the beating of the heart, which can then be used to judge the living body.
  • the 3 features of are output as a continuous signal rPPG.
  • the rPPG continuous signal is filtered, the Fourier transform is converted to the frequency domain, and the frequency domain analysis is performed to obtain the features, and the support vector machine model or the deep neural network model is used to classify the features to detect the living fingerprint.
  • the time-series deep neural network such as ConvLSTM
  • ConvLSTM ConvLSTM
  • three judgment results are obtained respectively according to the fingerprint image, according to the image under different brightness conditions and according to the rPPG signal, and the three judgment results are integrated to obtain the identification result.
  • the judgment result shows that the recognized object is a living body
  • the judgment result is 1
  • the judgment result shows that the recognized object is not a living body
  • the judgment result is 0
  • weights are set for the three judgment results, and the sum of each weight is equal to 1
  • the three judgment results are weighted, and a result value between 0 and 1 can be obtained.
  • the result value is greater than 0.5, the recognition result is that the recognized object is a living body; otherwise, the recognition result is a non-living body.
  • the method in an embodiment of the present application is used in a non-contact collection device as an example for description.
  • There are multiple ways to judge based on the fingerprint image that is, there are multiple judgment results, and the identification results are obtained by combining the multiple judgment results.
  • the fingerprint image of the recognized object is photographed first, and then based on the color fingerprint image of the recognized object, the image of the recognized object under different light source colors, the recognized object under different light source colors and the use of polarizers
  • the image of the recognized object, the infrared image of the recognized object, and the infrared temperature measurement result of the recognized object obtain five judgment results, and the five judgment results are combined to obtain the recognition result, which is described in detail below.
  • the steps of obtaining the judgment result according to the color fingerprint image of the recognized object may be the same as those in the above-mentioned embodiment, and the description will not be repeated here.
  • Light sources of different colors are used to illuminate the recognized object with light of different colors in turn, and images of the recognized object under the illumination of different colors are photographed, and the optical feature of the photographed image is recognized to obtain a judgment result.
  • the wavelengths of light of different colors are different, and objects of different materials will selectively absorb and reflect light of different wavelengths.
  • the light source emits light of a specific color, and after being reflected by the object, it appears in the imaging system that some wavelengths of light are lost.
  • Figure 6 shows the images of the same living finger under the illumination of four different colors of light. The images taken under light (wavelength 530nm), red light (wavelength 630nm) and white light, it can be seen that the imaging effect of the finger under different lighting is very different, this is because the skin on the finger is sensitive to light of different colors. The reflectivity is different.
  • the recognized object is a living body through methods such as support or deep neural network.
  • methods such as support or deep neural network.
  • a large number of non-contact images of living fingers and non-living fingers under the illumination of different colors of light sources can be collected, the reflectivity of different non-contact images under the illumination of each color can be calculated, and the difference in reflectivity can be used through the support vector machine. to determine whether the identified object is a living body.
  • different materials also exhibit different optical properties under different color light sources and polarizers.
  • a polarizer can be added in front of the camera lens of the camera that captures the image, and another polarizer is placed in front of the light source.
  • the polarization directions of the two polarizers are at a certain angle, and there can be multiple light sources. Different light sources correspond to polarizers with different polarization directions.
  • Figure 9 shows the fingerprint image of the same finger under the illumination of different color light sources when one polarizer is added to the camera lens and another polarizer is placed on the light source, the polarization of the polarizer on the camera lens and the polarizer on the light source directions cross.
  • Additional infrared fill light, etc. can be configured, and an IR-Cut filter can be configured on the camera.
  • an IR-Cut filter can be configured on the camera.
  • the skin on the surface of a human finger has a low reflectivity to infrared light, and infrared light can even penetrate the skin of the finger and illuminate the finger veins under the skin. Therefore, when the recognized object is a living body, an infrared image as shown in FIG. 8 can be obtained by shooting.
  • the optical properties of the finger surface of the infrared image can be analyzed, and the support vector machine or deep neural network can be used to determine whether the object to be recognized is a living body, or other methods can be used to detect the living body;
  • the characteristics of the finger veins in the image, and the finger vein characteristics are compared to determine whether the finger veins are the same as the finger vein characteristics of the pre-recorded living body.
  • the finger vein characteristics of the living body corresponding to the fingerprint image in the infrared image are obtained as the characteristics to be compared. , and compare the features of the finger veins in the infrared image with the features to be compared.
  • a series of infrared images of living fingers and non-living fingers are collected in advance.
  • y_i is used to represent whether the image is an image of a living body.
  • the infrared temperature sensor is used to photograph the palm area of the recognized object, and the temperature of the palm area is directly read. If the temperature is within the preset temperature range, it is determined that the recognized object is a living body, otherwise it is not a living body.
  • a single-point temperature measurement sensor can be used, or an area array temperature measurement sensor can be used.
  • an area array temperature measurement sensor is used, the average temperature of the palm area can be calculated to determine whether the recognized object is a living body or not.
  • the image of infrared temperature can be used in combination with deep neural network to accurately determine whether it is a living fingerprint.
  • the SVM/Boosting machine learning method can be used to combine the judgment results of multiple methods to obtain a more accurate final result. For example, if there are m methods, the decision values are used to represent the judgment results of each method, the decision values are Z_1,... , so that G(Z_1,..,Z_m) ⁇ y_i, so as to further improve the accuracy of living body detection by using various methods.
  • the fingerprint collection method of this embodiment may include:
  • an image acquisition step acquiring at least one image including a finger, wherein the image may include a finger image or a fingerprint image obtained according to the collected finger image;
  • the content of acquiring the finger image has been specifically described in the above-mentioned part of the non-contact fingerprint identification method, and only the difference is described here. For other content, please refer to the description in the above-mentioned part.
  • an image that has been photographed of the finger can also be obtained from the outside, which will be described in detail below.
  • the step of determining the incidental information determining the incidental information of the finger, wherein the incidental information includes at least one of the following information: joint finger information, finger position information, left and right hand information, mirror image information, and number of fingers information.
  • the incidental information is the information incidental to the image and is independent of the image.
  • the incidental information may be determined by the collection device itself, or acquired by the collection device from the user.
  • determining the incidental information by the collection end itself may include: the collection end determines the finger position information according to the collection device or user input, and determines the mirror image information according to whether the camera device used by the system is a front camera device or a rear camera device.
  • fa,fb,fc,fd...fj is the name of the fingerprint image
  • fa,fb,fc,fd...fj represents the order of the images at the time of submission
  • fa,fb,fc,fd...fj corresponds to Information other than the fingerprint image is incidental information.
  • the attached information includes a number immediately following the fingerprint image name, the number represents the corresponding finger position of the fingerprint image.
  • fa2 and fb3 represent a submission, which are finger positions 2 and 3 respectively; when the attached information does not include the fingerprint image name after the name
  • finger position information which means that the finger position corresponding to each submitted fingerprint image is unknown.
  • fa, fb, fc represent three fingerprints submitted at a time, no finger position.
  • the supplementary information includes "joint finger” after the name of the fingerprint image, it means that the supplementary information includes "is a joint finger", which means that the corresponding finger positions of the submitted fingerprint images are adjacent; when the supplementary information does not include the name of the fingerprint image
  • the following "joint finger” it means that the information on whether or not to join the finger is not included in the supplementary information, which means that the corresponding finger positions of the submitted fingerprint images may be adjacent or not.
  • the attached information When the attached information includes "left hand” or “right hand” after the name of the fingerprint image, it means that the designation of left and right hands is included in the attached information, which means that the submitted fingerprint image corresponds to the finger position of the left or right hand; when the attached information does not include fingerprints When "left-hand” or “right-hand” after the image name, it means that the attached information does not include left-hand or right-hand information, which means that the corresponding finger positions of the submitted fingerprint images may be left-handed or right-handed.
  • Incidental information includes, but is not limited to, the following forms:
  • Multiple fingers at a given position submit multiple fingers at a time, and all need to have a position, such as fa2, fb3, fa2, fb7, fa2, fb3, fc4, fd5;
  • Single-finger without specifying a single finger submit a single finger at one time without a finger, such as fa;
  • a ligature submit multiple fingers at one time without a ligature, such as fa, fb (ligature), fa, fb, fc, fd (ligature);
  • Left and right hand ligatures submit multiple fingers at one time, without finger position, it is ligature, and given left and right hand information, such as fa, fb (ligature, left hand), fa, fb, fc, fd (ligature , right hand).
  • setting at least one combination of fingerprint image and finger position based on the additional information includes: setting at least one fingerprint image and finger position based on the additional information and information determined from the fingerprint image. combination of bits. In this way, the possible combination of the fingerprint image and the finger position can be narrowed down according to the additional information and the information determined in the fingerprint image.
  • setting at least one combination of the fingerprint image and the finger position includes: determining the number of fingertip frames corresponding to the multiple fingertip regions. the center position of the center point; according to the clockwise or counterclockwise direction relative to the center position, determine the order of the plurality of fingertip frames; according to the order of the plurality of fingertip frames and the left and right hands in the accompanying information information to determine the combination of the fingerprint image and the finger position.
  • the incidental information determining step S112 may further include a finger position information determination step, that is, the step of detecting the finger position by a fingerprint collection device: determining the center positions of the plurality of fingertip frames corresponding to the plurality of fingertip regions ; Determine the order of the plurality of fingertip frames in a clockwise or counterclockwise direction relative to the center position; determine the order of the plurality of fingertip frames according to the order of the plurality of fingertip frames and the left and right hand information in the incidental information The combination of fingerprint image and finger position.
  • a finger position information determination step that is, the step of detecting the finger position by a fingerprint collection device: determining the center positions of the plurality of fingertip frames corresponding to the plurality of fingertip regions ; Determine the order of the plurality of fingertip frames in a clockwise or counterclockwise direction relative to the center position; determine the order of the plurality of fingertip frames according to the order of the plurality of fingertip frames and the left and right hand information in the incidental information The combination of fingerprint image
  • the step of determining the finger position information may include: detecting the region where each finger included in the image is located from the acquired image and determining a fingertip frame corresponding to each finger; determining each fingertip The center point of the frame and the center for the center point of each fingertip frame; respectively connect the center with the center point of each fingertip frame to obtain a number of connecting lines corresponding to the number of fingers; calculate the multiple connecting lines.
  • the included angle formed by the adjacent connecting lines in the line; the center points of the two fingertip boxes corresponding to the two connecting lines with the largest included angle are determined as the starting point and the ending point respectively; and the finger position information is determined according to the left and right hand information.
  • the fingers including the first joint in the image are identified, and the fingertip frame area of each finger is determined respectively.
  • four fingertip frames are taken as an example, and four fingertip frames are determined.
  • the center point of the fingertip frame and the center of the center points of the four fingertip frames are respectively connected with the center points of the four fingertip frames to obtain four connection lines;
  • the angle formed by the adjacent connection lines; the center points of the two fingertip boxes corresponding to the two connection lines with the largest included angle are determined as the starting point and the end point, respectively, and the finger position information is determined by combining the left and right hand information, and the starting point and the end point are respectively determined as 2 -3-4-5 fingers or 7-8-9-10 fingers.
  • the combination setting step first assumes a possible fingerprint image-finger correspondence situation based on the existing information, for example, given 3 fingerprints fafbfc, with the additional information "is a joint finger", assume that the fingerprint image - A combination of finger positions is that fafbfc corresponds to the index finger, middle finger and ring finger of the left hand respectively.
  • the extension to all other possible combinations under the condition of satisfying the incidental information includes: expanding to all other possible combinations under the condition that the condition of the incidental information is satisfied and the information determined in the fingerprint image is satisfied.
  • the fingerprint features of the finger are extracted from the fingerprint image, and the extracted fingerprint features are stored according to the combination of fingerprint image-finger position determined according to the accompanying information; it can be understood that if a fingerprint image corresponds to a finger position, the fingerprint feature extracted from the fingerprint image also corresponds to the finger position.
  • it may further include a normalization step of normalizing the acquired fingerprint images of the fingers respectively to obtain a normalized fingerprint image;
  • the ridge density of the fingerprint is obtained in the fingerprint parameter obtaining step, and in the normalizing step, according to the preset frequency, the ridge density is adjusted to Preset the ridge line density to obtain a normalized fingerprint image. From the above, compared with the image obtained by non-contact photography, the image obtained by contact photography does not need to obtain the information of finger position, finger outline and knuckle line, but directly extracts the fingerprint frequency/line density information for adjustment.
  • the finger contour such as the length of the knuckle line, the area of the contour line, etc.
  • the ridge density of the fingerprint belongs to the fine adjustment reference.
  • the images obtained by non-contact shooting can also be adjusted by using only the ridge line density as a reference.
  • a scale expansion step may also be included to perform at least one scale expansion on the acquired normalized fingerprint image to obtain fingerprint samples of at least one scale.
  • an image expansion step may also be included to expand the normalized fingerprint image to perform analog transformation, and obtain at least one fingerprint image of the finger.
  • the fingerprint recognition method of the embodiment of the present application does not rely on auxiliary hardware other than an image acquisition device, such as a mobile phone, by adjusting the size of the fingerprint image captured by the image acquisition device or collected by a contact collector, so that different batches of the same finger can be obtained by adjusting the size of the fingerprint image.
  • the collected fingerprint images are all scaled and adjusted to nearly the same size, keeping the relative distance and angle error of the fingerprint feature lines within a certain threshold, such as 5%, to support the subsequent processing of the fingerprint comparison system.
  • the fingerprint comparison method of this embodiment includes:
  • an image acquisition step acquiring at least one image including a finger; wherein, the image may include a finger image or a fingerprint image obtained according to the collected finger image;
  • the content of acquiring the finger image has been specifically described in the above-mentioned part of the non-contact fingerprint identification method, and only the difference is described here.
  • the description in the above-mentioned part please refer to the description in the above-mentioned part.
  • the incidental information includes at least one of the following information: continuous finger information, finger position information, left and right hand information, mirror image information, and finger quantity information.
  • Linked finger information, finger position information, left and right hand information, mirror image information, and number of fingers information include at least one of the following: single finger at a given finger position, multiple fingers at a given finger position, single finger without a given finger position, and no finger position given Linked fingers, left and right hands linked together. See S112 for the explanation of the incidental information, which will not be repeated here.
  • S133 a combination setting step, based on the additional information, setting at least one combination of fingerprint images and finger positions;
  • a possible fingerprint-finger correspondence situation is assumed first, for example, given 3 fingerprints abc, with the additional information "is a joint finger", a possible fingerprint is assumed -
  • the combination of finger positions is abc- left index finger, middle finger, ring finger.
  • the ring finger can be expanded into the following three hypotheses:
  • all the hypotheses that have been obtained so far are extended without changing other conditions, for example, the existing hypotheses abc - left hand ring finger, middle finger, index finger, in the unspecified In the case of left and right hands, it will be expanded to: abc - right ring finger, middle finger, index finger.
  • all hypotheses that have been obtained so far are extended without changing other conditions, eg, based on existing mirroring information. For example, it has been assumed that abc - right ring finger, middle finger, index finger. If it is not specified whether the fingerprint data is mirrored or not, it will be expanded to: abc - right ring finger, middle finger, index finger. The finger position is assumed to be unchanged, but all fingerprint images are mirrored left and right.
  • the incidental information determining step S132 may further include a finger position information determining step: dividing at least one finger included in the image into separate areas; determining the center position of each area; clockwise or counterclockwise The hour hand determines the order of the regions; the finger information is determined according to the order.
  • the fingers including the first joint in the image are identified, and the fingertip frame area of each finger is determined respectively.
  • the center position of each fingertip frame, and the order of the frame is determined in a clockwise/counterclockwise direction relative to the center, and the finger position information is determined in conjunction with the left and right hand information.
  • the step S132 of determining the incidental information may further include setting at least one combination of the fingerprint image and the finger position, including: setting at least one fingerprint image and the fingerprint image based on the incidental information and the information determined from the fingerprint image. Combination of finger positions. In this way, the possible combination of the fingerprint image and the finger position can be narrowed down according to the additional information and the information determined in the fingerprint image.
  • setting at least one combination of the fingerprint image and the finger position includes: determining the number of fingertip frames corresponding to the multiple fingertip regions. Center position; according to the clockwise or counterclockwise direction relative to the center position, determine the order of the plurality of fingertip frames; according to the order of the plurality of fingertip frames and the left and right hand information in the incidental information, determine The combination of the fingerprint image and the finger position.
  • the step of determining the finger position information may include: detecting the region where each finger included in the image is located from the acquired image and determining a fingertip frame corresponding to each finger; determining each fingertip The center point of the frame and the center for the center point of each fingertip frame; respectively connect the center with the center point of each fingertip frame to obtain a number of connecting lines corresponding to the number of fingers; calculate the multiple connecting lines.
  • the included angle formed by the adjacent connecting lines in the line; the center points of the two fingertip boxes corresponding to the two connecting lines with the largest included angle are determined as the starting point and the ending point respectively; and the finger position information is determined according to the left and right hand information.
  • the fingers including the first joint in the image are identified, and the fingertip frame area of each finger is determined respectively.
  • four fingertip frames are taken as an example, and four fingertip frames are determined.
  • the center point of the fingertip frame and the center of the center points of the four fingertip frames are respectively connected with the center points of the four fingertip frames to obtain four connection lines;
  • the angle formed by the adjacent connection lines; the center points of the two fingertip boxes corresponding to the two connection lines with the largest included angle are determined as the starting point and the end point, respectively, and the finger position information is determined by combining the left and right hand information, and the starting point and the end point are respectively determined as 2 -3-4-5 fingers or 7-8-9-10 fingers.
  • a normalization step is further included, and the acquired images of the fingers are respectively normalized to obtain a normalized fingerprint image;
  • the steps include: judging whether it is a photographing acquisition method; if so, extracting one of finger position information, finger contour information, and knuckle line information of the fingerprint data; extracting fingerprint frequency according to one of the finger position information, finger contour information, and knuckle line information information; the fingerprint image is normalized according to the fingerprint frequency information.
  • the finger outline and knuckle line recognition are provided by the deep learning model, from which the approximate size of the finger that needs to be scaled can be estimated.
  • the normalizing step may further include normalizing the ID card fingerprint image and the imprint fingerprint image.
  • ID card fingerprint images the area is relatively small, and generally does not contain finger shape and knuckle line information.
  • the fingerprint line density can be used for normalization; for the fingerprint image collected by the contact collector, the frequency of lines can be measured. , thereby maintaining compatibility with such images for normalization.
  • the method for extracting fingerprint frequency information may include: further performing local scaling, depending on the calculation of the fingerprint frequency, the specific method is to take a local fingerprint image within a sliding window; using the fingerprint image The translation invariance between the approximately parallel ridges is calculated, and the minimum translation invariance distance is calculated, thereby estimating the fingerprint frequency.
  • the content of the normalization process has been specifically described in the above section, and for details, please refer to the description in the above section.
  • a scale expansion step may also be included to perform at least one scale expansion on the obtained normalized fingerprint image to obtain at least one scale fingerprint sample, so that at the same time, in each person-time, each finger Corresponding to multiple fingerprint images, the number is determined by the proportional strategy and the expansion plan respectively;
  • an image expansion step may also be included to expand the image of the finger to perform analog transformation, and obtain at least one fingerprint image of the finger, and expand the fingerprint image obtained in the form of each image.
  • the algorithm simulates a flat fingerprint from a photographed fingerprint, and uses different expansion parameters to obtain multiple fingerprints, the number of which is determined by the number of fingerprints and the amount of given information.
  • the fingerprint comparison step extracting the fingerprint feature of the finger and comparing it with the existing fingerprint data according to at least one combination method with the incidental information
  • the fingerprint data is acquired by at least one of the following acquisition methods:
  • Fingerprint collectors There are various forms of collectors. The collection that can be completed at one time includes single-finger, multi-finger, one-hand four-finger, two-handed thumb, etc. The technology of the collector can be roughly divided into contact type Collector & non-contact collector;
  • Obtaining fingerprints by direct shooting For example, generally shooting with a mobile device, it can also be generalized to use a mobile device/fixed device to shoot a finger to obtain data;
  • Fingerprint data reading such as ID card readers; reading devices for ID documents such as social security cards, driver's licenses, etc.
  • At least one of the following information of the fingerprint data to be compared is the same or different: acquisition method information, pixel point information, data format information, and of course other different information may also be included, such as: Fingerprint, which is generally the output of the fingerprint collector, or read from the medium that stores the printed fingerprint, and is mostly a black and white image. The difference between different printed fingerprint images is mainly in DPI, pixel, color, grayscale, etc.
  • fingerprint features it can be read from the fingerprint feature file from the medium storing the fingerprint features, of course, it can also be obtained by some fingerprint collection devices The fingerprint features are extracted after the fingerprints are collected.
  • the result determination step S135 further includes:
  • the step of obtaining the score is to obtain the highest score obtained after comparing the fingerprint feature of at least one finger with the existing fingerprint data in at least one combination mode as the comparison score of the finger.
  • obtaining at least one combination mode The score corresponding to each combination mode, the score is determined according to the fingerprint feature of at least one finger under the combination mode and the additional information according to the existing fingerprint data described in each combination mode; The highest score among the above scores is used as the comparison score of the finger.
  • the comparison scores of different fingers are summarized, and the weighted square sum of the comparison scores of the multiple fingers in at least one combination mode is calculated as the final comparison score.
  • the maximum value of the final alignment scores of at least one combination is used as the final alignment result.
  • an embodiment of the present application provides a fingerprint collection and comparison device, which can be specifically applied to various electronic terminal equipment, and is characterized in that , the fingerprint collection and comparison device may include modules corresponding to the steps, which will not be repeated here.
  • the present application discloses a set of fingerprint area detection methods to deal with this technical problem.
  • the fingerprint area detection method of an embodiment of the present application mainly includes the following steps:
  • S141 an image acquisition step, performing non-contact photography on the subject's hand and acquiring an image of at least one finger including knuckles;
  • the content of acquiring the finger image has been specifically described in the above-mentioned part of the non-contact fingerprint identification method, and only the difference is described here.
  • the description in the above-mentioned part please refer to the description in the above-mentioned part.
  • the direction adjustment step may include: identifying the end points of the knuckles in the image; determining the direction of the knuckle line according to the end points of the knuckles; adjusting the image according to the direction of the knuckle line, so that the adjusted The knuckle lines in the image are oriented, for example, in a direction perpendicular to a preset direction.
  • a fingertip image including the first finger joint in the image is identified, the end point of the first finger joint is identified on the fingertip image, the direction of the knuckle line is determined according to the end point of the first finger joint, and the image is adjusted according to the direction of the knuckle line , so that the knuckle line orientation remains horizontal in the adjusted image.
  • the direction adjustment step may further include: dividing at least one finger including knuckles in the image into separate areas; determining the direction of the divided separate areas; adjusting the direction of the separate areas so that the finger faces the predetermined direction
  • the direction is set, for example, four fingers are divided into four separate areas, the pointing direction of the fingers in the divided separate areas is determined, and the pointing direction of the fingers in each separate area is directed to the vertical upward direction.
  • the direction adjustment step may further include: determining the direction of the outer edge of at least one finger including knuckles in the image; adjusting the direction of the outer edge to adjust the direction of the finger so that the finger faces a preset direction, For example, the outer edges of the four fingers are respectively determined, the direction of the formed outer edge image is determined, and the direction of each outer edge is directed to the vertical upward direction.
  • the background processing step includes: identifying the edge of the finger and the background on the image and forming a polygon with the edge; obtaining an image inside the polygon as a foreground, and removing the background outside the polygon.
  • the background processing step may further include: dividing at least one finger including knuckles in the image into separate regions and acquiring finger images in the separate regions, and removing the background outside the finger images in the separate regions , for example, divide four fingers into four separate areas, obtain the image of the finger in each separate area, determine the background outside the image of the finger in the divided separate area, and divide the image of the finger outside each separate area. Background removal.
  • in the direction adjustment step at least one finger including a knuckle in the image is divided into separate regions, and after the image of the fingertip in the separate region is acquired and the direction is adjusted, the parts other than the image of the fingertip in the separate region are removed. background.
  • the step of determining the finger position may include: dividing at least one finger including a knuckle in the image into separate regions; determining the center position of the region; determining the order of the regions clockwise or counterclockwise; The order determines the pointing information. Referring to Fig. 9, for example, a picture of a finger including the first joint in the image is identified and a fingertip frame is formed for each finger,
  • a scale obtaining step is further included, and the scale information corresponding to the physical world is obtained by using the width of the finger.
  • the present application discloses a set of fingerprint normalization methods to deal with this technical problem.
  • the fingerprint identification method according to the embodiment of the present application is described below, including the following steps.
  • S161 acquiring step, acquiring an image including a fingerprint.
  • various image acquisition devices can be used in the embodiments of the present application, including but not limited to a camera device, a mobile phone provided with a camera device, a collector that collects fingerprints by contact, and the like.
  • the present embodiment can acquire images including fingerprints in various forms.
  • the content of acquiring the fingerprint image has been specifically described in the above-mentioned part of the non-contact fingerprint identification method, and only the difference is described here. For other content, please refer to the description in the above-mentioned part.
  • the determination step is to determine whether the image is an image obtained by non-contact photographing of the fingerprint.
  • the embodiments of the present application can determine whether the image is an image obtained by non-contact photography of a fingerprint according to the attributes of the image; or determine whether the image includes the outer contour of the finger and/or the knuckle line according to the deep learning model. Steps; if the outer contour and/or the knuckle line of the finger is included in the image, the image is an image obtained by non-contact photographing of the fingerprint; if the outer contour and/or the knuckle line of the finger is not included in the image, the image is Image obtained by contact acquisition of fingerprints.
  • deep learning is a general term for a type of pattern analysis method.
  • a neural network system based on convolution operation namely convolutional neural network (CNN).
  • CNN convolutional neural network
  • Auto-encoding neural network based on multi-layer neurons, including auto-encoder and sparse coding.
  • Pre-training is performed in the form of a multi-layer self-encoding neural network, and then combined with the discriminant information to further optimize the deep belief network (DBN) of the neural network weights.
  • DBN deep belief network
  • Deep learning combines low-level features to form more abstract high-level representation attribute categories or features to discover distributed feature representations of data.
  • the computation involved in producing an output from an input can be represented by a flow graph: a flow graph is a graph capable of representing computations, in which each node represents a basic computation as well as a computation The value of the calculation is applied to the values of the child nodes of this node.
  • the fingerprint parameter acquisition step when it is determined that the image is obtained by non-contact photographing of the fingerprint, the size and ridge density of the fingerprint are acquired.
  • the embodiments of the present application may include: obtaining two endpoints of the knuckle line according to the knuckle line model, and obtaining the distance between the two endpoints; or obtaining the two endpoints of the knuckle line according to the knuckle line model , according to the finger contour model, obtain a plurality of boundary points on the edge of the finger; smoothly connect two end points and a plurality of boundary points in turn to form the contour of the finger; and calculate the area of the contour to obtain the area; wherein, in the normalization step Among them, adjusting the size to the preset size includes: adjusting the length to the preset length; or adjusting the distance to the preset distance; or adjusting the area to the preset area.
  • FIG. 10 is a schematic diagram of the outline of an embodiment of the present application.
  • the contour line may be surrounded by a plurality of boundary points and knuckle lines of the edge of the finger.
  • point A and point B are the two end points of the knuckle line, respectively.
  • the normalization step according to the preset standard, adjust the size to the preset size, adjust the ridge density to the preset ridge density, and obtain a normalized fingerprint image.
  • the embodiments of the present application may include the steps of dividing the area where the fingerprint is located into a plurality of sub-areas; the step of separately acquiring the fingerprint direction and ridge density of each sub-area in the plurality of sub-areas; The steps of obtaining the minimum translation invariant distance of each sub-region respectively; according to the minimum translation invariant distance, adjust the ridge density of each sub-region respectively, so that the ridges of each adjacent sub-region are The difference value of the density is smaller than the preset difference value; and the steps of obtaining the preset ridge line density according to the preset frequency.
  • the embodiment of the present application can also take a local fingerprint image inside a sliding window, and use the translation invariance between approximately parallel ridges to calculate the minimum translation invariant distance, thereby estimating the fingerprint frequency.
  • S165 a fingerprint identification step, performing fingerprint identification based on the normalized fingerprint image obtained in the normalization step.
  • the embodiments of the present application may further include the step of extracting a fingerprint region from the image to exclude background regions in the image, wherein the fingerprints in the fingerprint region have a substantially uniform ridge density.
  • the embodiments of the present application may further include positioning the finger position of the finger by combining the user's input of the left and right hands and the deep learning model to detect the finger position frame.
  • FIG. 11 schematically shows a schematic diagram of a non-contact fingerprint acquisition device fixed on a wall.
  • the non-contact acquisition includes: a casing 1 , an image acquisition device 2 and a processing device 3 , wherein the casing 1 includes a first component part 11 and a second component part 12 which are connected to each other in an L shape.
  • the image acquisition device 2 is arranged on the casing 1 and is used for taking fingerprint images.
  • the photographing area of the image acquisition device 2 is located on the side of the first component part 11 close to the second component part 12; the processing device 3 is arranged in the casing 1, It is used to perform recognition processing on the images captured by the image acquisition device 2 .
  • the housing 1 can be made of non-transparent materials, the shooting direction of the image acquisition device can be directed vertically downward, and the image acquisition device 2 is used for shooting non-contact fingerprint images.
  • the second composition The side of the part 12 away from the first component part 11 can be used for fixing, for example, on a wall as shown in FIG. 11 . Since the first component part 11 and the second component part 12 form an L-shaped structure, and the shooting area of the image capture device 2 is located on the concave side of the L-shape, the first component part 11 and the second component part 12 can block most of the stray light , thereby reducing the interference of stray light on the acquisition of non-contact fingerprint images.
  • the housing 1 in this embodiment can block stray light in at least two directions to ensure the quality of the images collected by the image collection device 2 .
  • the housing 1 further includes: a third component part 13 ; in this embodiment, the first component part 11 , the second component part 12 and the third component part 13 Connected in sequence, and the first component part 11 and the third component part 13 are relatively spaced apart, the first component part 11, the second component part 12 and the third component part 13 surround to form a collection space, and the shooting area of the image acquisition device 2 is located in the collection space. within the space.
  • the first component part 11 , the second component part 12 and the third component part 13 form a concave structure.
  • the photographing direction of the image capture device 2 faces the third component part 13 , At this time, the third component 13 blocks the light directly entering the image capturing device 2 , thereby greatly reducing the stray light entering the image capturing device 2 and further improving the clarity of the image captured by the image capturing device.
  • the non-contact image capture device further includes: an illumination device 4 , which is arranged in the housing 1 and is used for illuminating when the image capture device 2 shoots.
  • the illuminating device 4 may include one or more light sources, and the illuminating device 4 illuminates the finger photographed by the image capturing device 2 , so that the brightness of the photographed finger can be significantly improved to improve the contrast of the fingerprint.
  • the lighting device 4 includes: at least two lighting components 41 ;
  • the colors of the light are different, so that the image acquisition device can capture fingerprint images under different light colors. Specifically, during fingerprint collection, different colors of light can be emitted in sequence, so that fingerprint images under different colors of light can be obtained.
  • This method can effectively prevent fake fingers, because the human body and rubber and other materials are in different colors.
  • There are differences in the reflectance under the illumination of different colors so the reflectance of the photographed finger can be calculated according to the images under the illumination of different colors, so as to judge whether the photographed finger is a human finger or a fake finger.
  • the non-contact fingerprint acquisition device in this embodiment further includes: an optical path adjustment device 5 , which is arranged on the side of the housing 1 close to the shooting area of the image acquisition device 2 ,
  • the lens of the image acquisition device 2 faces the optical path adjustment device 5
  • the optical path adjustment device 5 is used to change the shooting direction of the image acquisition device 2
  • the image acquisition device 2 shoots a fingerprint image of the shooting area through the optical path adjustment device 5 .
  • the reason for setting the optical path adjustment device 5 is that when the distance between the image acquisition device 2 and the finger is small, the captured fingerprint image is prone to distortion.
  • the optical path adjustment device 5 includes: a reflector; the reflector is arranged on the side of the housing 1 close to the shooting area of the image acquisition device 2, and the plane where the reflector is located is related to the image acquisition There is a preset angle between the shooting directions of the device 2 .
  • the image of the finger is reflected into the lens of the image acquisition device 2 through the mirror, and the shooting area of the image acquisition device 2 can be adjusted by adjusting the shooting direction of the image acquisition device and the preset angle between the mirror.
  • the non-contact fingerprint collection device further includes: a structured light projection device 6 , which is arranged on the housing 1 and is used for projecting structured light when the image capture device 2 shoots, So that the image acquisition device 2 captures the structured light image.
  • a structured light projection device 6 By adding the structured light projection device 6, the user's finger image with structured light stripes can be captured.
  • the processing device 3 processes the image, the finger can be 3D modeled according to these stripes, and then the identified and cut out The finger image is expanded to obtain a finger image that is closer to pressing.
  • the non-contact fingerprint collection device further includes: a time-of-flight device 7; the time-of-flight device 7 is used to emit infrared light pulses to the object to be photographed, and The returned optical signal calculates the depth information of the object to be photographed; the image acquisition device 2 is used for focusing when photographing the fingerprint image of the object to be photographed according to the depth information.
  • the time-of-flight ranging method is adopted in this embodiment, and the time-of-flight device 7 can transmit continuous infrared light pulses of a specific wavelength to the object to be photographed (such as a finger), and receive the to-be-to-be-photographed through the sensor on the time-of-flight device 7.
  • the optical signal returned by the photographed object is used to calculate the round-trip flight time or phase difference of the light to obtain the three-dimensional depth information of the object to be photographed.
  • the camera often takes a long time to focus, and the user experience is poor. Therefore, this embodiment A time-of-flight device (TOF module) is added to measure the distance from the image acquisition device to the finger, and then directly specify the distance when focusing to achieve the effect of fast focusing.
  • TOF module time-of-flight device
  • the image acquisition device 2 includes: multiple cameras, at least two of the multiple cameras are focused at different positions; in this way, images at multiple positions can be captured simultaneously, which expands the space where a finger can be placed. area, and multiple cameras can improve the depth of field, if the depth of field of each camera is k cm, by using n cameras, a system with a total depth of field of k*n cm depth of field can be assembled. Choose an image with the highest quality.
  • the imaging system of the liquid lens of the image acquisition device can change the focal plane within milliseconds, providing a sharp image regardless of the distance of the object from the camera, which makes the integrated liquid lens necessary to focus at multiple distances Ideal for shooting situations, the liquid lens can take 250 images per second step, and the required energy follows the tip.
  • the liquid lens consists of a pair of water droplets, which are exposed to high-frequency sound waves and vibrate back and forth, which changes the focal length of the lens.
  • the software can automatically step the pictures within the focal distance and discard the pictures outside the focal distance.
  • the image capture device 2 has a Fraser lens
  • the Fraser lens has a great depth of field, and can focus on extremely far and extremely close objects at the same time for shooting, and the Fraser lens uses a larger imaging lens
  • the wide-angle lens projects the effect with a large depth of field on a film, and then uses a zoom lens to capture the image on the film to achieve zoom and control the aperture.
  • the Fraser lens achieves far more than ordinary lenses. depth of field.
  • the image acquisition device 2 includes: a light field camera.
  • the light field camera can capture information about the direction of the light in the scene and record the data of the beams in all directions, so it can focus on any depth in the photographed photos, and then focus through the software according to the actual picture needs later to get a clearer picture effect .
  • a non-contact fingerprint collection method is also proposed in the embodiment of this application, and the method in this embodiment includes:
  • the acquisition step a non-contact image of the to-be-recognized fingertip of the finger is photographed, and the to-be-recognized fingertip image is a fingertip image with a knuckle line, as shown in the outer frame of FIG. 2 .
  • any one of the non-contact fingerprint collection devices proposed in this application is used to collect a fingerprint image.
  • the finger needs to be recognized, and the process of recognizing the finger can be as described in any of the embodiments of the application. finger recognition steps.
  • one or more fingers are captured in the captured fingerprint image. Therefore, in order to facilitate the separate processing of the fingerprint images of each finger, it is necessary to firstly segment the fingers in the captured and processed fingertip images to obtain each finger. Fingerprint image.
  • the processing step specifically adopts the fingerprint acquisition step in any embodiment of the present application.
  • the fingerprint images corresponding to each finger are stored only when the conditions are met, and the images of each finger are not stored when the conditions are not met.
  • the preset conditions can be, for example, collecting The definition of the fingerprint image meets the requirements, the collected fingerprint image is confirmed to be the fingerprint image of the living body, and the collected fingerprint image is stored when the conditions are met, so that it can be used for fingerprint comparison and other operations. Fingerprint image.
  • the capturing step may include capturing multiple sets of fingerprint images to be identified with different focuses.
  • the fingers of the human body are not in the same plane, and the fingers themselves also have ups and downs. If only one fingerprint image is collected, the resolution of different fingers may be insufficient. By collecting multiple sets of fingerprints with different focus to be identified image, which can ensure that the fingerprint image of each finger is clear enough.
  • the relative aperture distance of the lens of the camera is reduced when the fingerprint image to be recognized is photographed, and the light is supplemented when photographing.
  • the relative aperture distance of the lens is roughly inversely proportional to the depth of field.
  • the image of the fingertip to be identified is a structured light image; the storing step further includes: establishing a corresponding 3D fingerprint model according to the fingerprint image when conditions are met, and overlaying the fingerprint image on the corresponding 3D fingerprint model, The 3D fingerprint model is expanded to obtain the corresponding 2D fingerprint image, and the 2D fingerprint image is stored.
  • the storing step further includes: when conditions are met, establishing a corresponding 3D fingerprint model according to the fingerprint image, and expanding the fingerprint image according to the part corresponding to the fingerprint image in the 3D fingerprint model to obtain the corresponding fingerprint image. 2D fingerprint image, store the 2D fingerprint image.
  • the non-contact fingerprint collection device used in this embodiment has a structured light projection device, which captures a structured light image.
  • the structured light technology can be used to obtain 3D image information and improve the accuracy of 3D imaging. Therefore, in this embodiment, the finger can be obtained.
  • the 3D information is recorded in the stored 2D image, which greatly enriches the information in the fingerprint image, which is beneficial to improve the fingerprint comparison accuracy in the later stage.
  • the processing steps include: identifying the first knuckle line of the fingertip image to be recognized, and obtaining only the fingers above the first knuckle line from the to-be-recognized fingertip image according to the first knuckle line. Fingertip image of the tip area; the finger foreground image is obtained as the fingerprint image from the fingertip image that contains only the fingertip area above the first knuckle line.
  • the finger includes a plurality of knuckles, the knuckle line located closest to the end of the finger is the first knuckle line, and the fingerprint is located at the fingertip of the finger, so only the image of the fingertip needs to be processed, and no other processing of the finger is required. Collecting and processing part of the fingertip image can reduce the amount of data processing in the later stage and improve the response speed. Similarly, there is no fingerprint information in the background of the fingertip image, so only the foreground image of the finger needs to be processed.
  • obtaining the finger foreground image from the fingertip image may include: identifying the edge of the finger and the background in the fingertip image, and obtaining the finger foreground image from the fingertip image according to the edge.
  • identifying the first knuckle line of the to-be-identified fingertip image may include: identifying the fingertip including the first knuckle in the to-be-identified fingerprint image; identifying the knuckle endpoint of the first knuckle; Determine the knuckle line of the first knuckle.
  • the fingertip is firstly found in the fingerprint image to be recognized. Because the fingertip is located at the end of the finger, the area corresponding to the fingertip can be found relatively quickly. Since the finger has multiple traces such as fingerprint lines , it is difficult to find the first knuckle line directly.
  • this embodiment can improve the calculation speed, and because there are fewer feature points on the finger contour, the accuracy can be improved.
  • the method may further include: adjusting the pointing of the finger in the fingertip image according to the first knuckle line.
  • the position of the finger is adjusted, for example Let the finger point in a preset direction, for example, the preset direction can be directly above the image, at this time, the finger pointing can be adjusted by adjusting the first knuckle line to be parallel to the horizontal direction in the image.
  • the processing step may further include: determining the relative positional relationship of each finger in the fingerprint image to be recognized; and determining the finger name of each finger according to the relative positional relationship and the left and right hand information of the fingers.
  • the storing step may further include: storing the name of the finger corresponding to the fingerprint image, that is, the finger position. Specifically, there are multiple fingers in the fingerprint image captured in this embodiment, and the shapes of the left and right hands are obviously different, so it can be directly determined according to the captured fingerprint image whether the captured fingerprint image is the left hand or the right hand. The changes are different and have a relative positional relationship with each other.
  • the relative positional relationship may include, for example, the relative position of the head of each finger.
  • each finger is the index finger, the middle finger or the ring finger.
  • the thumb is obviously not in line with the position of the other fingers. It can be clearly determined which one is the thumb according to the relative position of the thumb and the other fingers, and according to the position of the thumb Determine whether it is the left hand or the right hand, and then determine the name of each finger according to the order of the remaining four fingers. For another example, when taking an image of four fingers without the thumb, since the middle finger is the highest, the middle finger can be easily determined. , according to the number of fingers on the left and right of the middle finger, which one is the index finger can be determined, or the finger names of the other fingers can be determined according to the position of the shortest little finger.
  • the judging step may further include: determining whether the finger is a living object according to the fingerprint image, and if the finger is not a living object, determining that the condition is not met.
  • the method recorded in the living body identification step in any embodiment of the present application may be adopted, and details are not described here.
  • the implementation of the apparatus since it basically corresponds to the implementation of the method, it is sufficient to refer to the partial description of the implementation of the method for related parts.
  • the device implementations described above are merely illustrative, wherein the modules described as separate modules may or may not be separate. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this implementation manner. Those of ordinary skill in the art can understand and implement it without creative effort.
  • each block in the block diagrams of the accompanying drawings may represent a module, a portion of which may contain one or more executable instructions for implementing the specified logical function, the modules are not necessarily in order Execute in sequence.
  • the modules and functional units in the device implementations in this application may be integrated into one processing module, or each unit may exist physically alone, or two or more modules or functional units may be integrated into one module.
  • Each of the above-mentioned integrated modules can be implemented in the form of hardware, and can also be implemented in the form of software function modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.
  • the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, and the like.
  • FIG. 15 shows a schematic structural diagram of an electronic device (eg, a terminal device or a server in FIG. 1 ) 500 suitable for implementing the embodiments of the present application.
  • the terminal device in the embodiments of the present application may be various terminal devices in the above-mentioned system.
  • the electronic device shown in FIG. 15 is only an example, and should not impose any limitations on the functions and scope of use of the embodiments of the present application.
  • the electronic device 500 may include a processing device (eg, a central processing unit, a graphics processor, etc.) 501 for controlling the overall operation of the electronic device.
  • the processing device may include one or more processors to execute instructions to perform all or part of the steps of the above-described methods.
  • the processing device 501 may also include one or more modules for processing interactions with other devices.
  • the storage device 502 may be configured to store various types of data, and the storage device 502 may include various types of computer-readable storage media or combinations thereof, such as electrical, magnetic, optical, electromagnetic, infrared, or A semiconductor system, apparatus or device, or any combination of the above. More specific examples of computer readable storage media may include, but are not limited to, electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable Programmable read only memory (EPROM or flash memory), fiber optics, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
  • a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
  • the sensor device 503 which may be configured to sense prescribed measured information and convert it into a usable output signal according to a certain rule, may include one or more sensors.
  • it may include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor, etc., for detecting changes in the electronic device's open/closed state, relative positioning, acceleration/deceleration, temperature, humidity, and light, etc.
  • the processing device 501 , the storage device 502 and the sensor device 503 are connected to each other by a bus 504 .
  • An input/output (I/O) interface 505 is also connected to bus 504 .
  • the multimedia device 506 may include input devices such as a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, etc. to receive input signals from the user. Various input devices may cooperate with various sensors of the sensor device 503 to complete, for example, gestures. Operation input, image recognition input, distance detection input, etc.; the multimedia device 506 may also include an output device such as a liquid crystal display (LCD), a speaker, a vibrator, and the like.
  • input devices such as a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, etc.
  • Various input devices may cooperate with various sensors of the sensor device 503 to complete, for example, gestures. Operation input, image recognition input, distance detection input, etc.; the multimedia device 506 may also include an output device such as a liquid crystal display (LCD), a speaker, a vibrator, and the like.
  • LCD liquid crystal display
  • Power supply device 507 which may be configured to provide power to various devices in the electronic device, may include a power management system, one or more power supplies, and components that distribute power to other devices.
  • the communication means 508 may allow the electronic device 500 to communicate wirelessly or by wire with other devices to exchange data.
  • the above-mentioned devices can also be connected to the I/O interface 505 to realize the application of the electronic device 500 .
  • Figure 15 illustrates an electronic device having various means, it should be understood that not all of the illustrated means are required to be implemented or available. More or fewer devices may alternatively be implemented or provided.
  • embodiments of the present application include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated in the flowchart.
  • the computer program may be downloaded and installed from a network via a communication device, or installed from a storage device.
  • the processing device When the computer program is executed by the processing device, the above-described functions defined in the methods of the embodiments of the present application are executed.
  • a machine-readable medium may be a tangible medium that may contain or store the program for use by or in connection with the instruction execution system, apparatus or device.
  • the computer-readable medium described above in the present application may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two.
  • a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code therein. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device .
  • Program code embodied on a computer readable medium may be transmitted using any suitable medium including, but not limited to, electrical wire, optical fiber cable, RF (radio frequency), etc., or any suitable combination of the foregoing.
  • the above-mentioned computer-readable medium may be included in the above-mentioned electronic device; or may exist alone without being assembled into the electronic device.
  • Computer program code for performing the operations of the present application may be written in one or more programming languages, including but not limited to object-oriented programming languages—such as Java, Smalltalk, C++, and This includes conventional procedural programming languages - such as the "C" language or similar programming languages.
  • the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server.
  • the remote computer may be connected to the user's computer through any kind of network, or may be connected to an external computer (eg, using an Internet service provider to connect over the Internet).
  • each block in the flowchart or block diagrams may represent a module, segment, or portion of code that contains one or more logical functions for implementing the specified functions executable instructions.
  • the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented in dedicated hardware-based systems that perform the specified functions or operations , or can be implemented in a combination of dedicated hardware and computer instructions.
  • the units involved in the embodiments of the present application can be implemented in software or hardware. Among them, the name of the unit does not constitute a limitation of the unit itself under certain circumstances.
  • exemplary types of hardware logic components include: Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), Systems on Chips (SOCs), Complex Programmable Logical Devices (CPLDs) and more.
  • FPGAs Field Programmable Gate Arrays
  • ASICs Application Specific Integrated Circuits
  • ASSPs Application Specific Standard Products
  • SOCs Systems on Chips
  • CPLDs Complex Programmable Logical Devices
  • a non-contact fingerprint identification method comprising:
  • Comparing the fingerprint image with the fingerprint database image to obtain a comparison result including:
  • the incidental information includes at least one of link finger information, finger position information, left and right hand information, mirror image information, and number of fingers information;
  • comparing the fingerprint image with the fingerprint database image to obtain a comparison result including:
  • determining the fingerprint image-finger combination including:
  • At least one fingerprint image-finger combination that satisfies the condition of the incidental information is determined.
  • determining at least one fingerprint image-finger combination that satisfies the condition of the incidental information includes: based on the incidental information, determining all possible fingerprint image-finger combinations that satisfy the condition of the incidental information.
  • Item 4 The contactless fingerprint identification method according to any one of Items 2-3, wherein,
  • the step of determining at least one combination of fingerprint image and finger position includes:
  • the combination of the fingerprint image and the finger position is determined according to the sequence of the plurality of fingertip frames and the left and right hand information in the incidental information.
  • determining the order of the plurality of fingertip frames comprises:
  • the order of the plurality of fingertip frames is determined in the direction from the start point to the end point.
  • Item 6 The contactless fingerprint identification method according to any one of Items 1 to 5, wherein the image processing on the one or more fingertip images includes preprocessing, normalization, and image expansion. At least one, wherein the preprocessing includes background removal, or direction adjustment and background removal: the normalization includes at least one of coarse normalization and fine normalization; the image expansion includes zoom-in, zoom-out, at least one of image flattening;
  • the direction adjustment includes: adjusting the direction according to the direction of the first phalangeal line of the finger; adjusting the direction according to the direction of a separate area; or performing the direction adjustment according to the direction of at least one outer edge of the finger including the knuckle direction adjustment;
  • the removing the background includes: removing the background according to the foreground area enclosed by the fingertip outline and the first knuckle line; or removing the background outside the finger area in the separate area;
  • the first knuckle line is determined by a knuckle line model
  • the fingertip contour line is formed by acquiring multiple boundary points of the finger edge through the contour line model and connecting the boundary points.
  • the separate area is obtained by segmenting at least one of the fingers in the image that includes a knuckle.
  • Item 7 The contactless fingerprint identification method according to Item 6, wherein the coarse normalization comprises:
  • the fine normalization includes: detecting the ridge density of a plurality of sub-regions included in the fingerprint image; and
  • the overall frequency is calculated according to the ridge line density of the plurality of sub-regions, and the overall frequency is adjusted to the overall preset value.
  • Item 8 The non-contact fingerprint identification method according to Item 7, wherein detecting the ridge line density of the plurality of sub-regions included in the fingerprint image comprises:
  • the ridge density of each of the plurality of sub-regions is calculated according to the minimum translation-invariant distance.
  • Item 9 The non-contact fingerprint identification method according to any one of Items 6-8, wherein the enlarging in the image expansion comprises scaling up the normalized fingerprint image at least once to obtain an enlarged fingerprint image;
  • the reduction in the image expansion includes scaling down the normalized fingerprint image at least once to obtain a reduced fingerprint image, and the flattening in the image expansion includes: establishing a corresponding 3D fingerprint model according to the fingerprint image, and according to the 3D fingerprint image.
  • the part corresponding to the fingerprint image in the model expands the fingerprint image to obtain a 2D fingerprint image corresponding to the fingerprint image.
  • Item 10 The contactless fingerprint identification method according to any one of Items 2-9, wherein obtaining the comparison result comprises:
  • the weighted summation of the comparison scores of the fingerprint images corresponding to the plurality of finger positions under the combination is used as the final comparison score under the combination;
  • the comparison result is determined according to the maximum value among the final comparison scores corresponding to each combination.
  • Item 11 The non-contact fingerprint identification method according to Item 10, wherein the comparison score of the fingerprint image corresponding to each finger position is the fingerprint image corresponding to the finger position, the enlarged fingerprint image, and the reduced fingerprint image.
  • the highest score obtained by comparing at least one of the 2D fingerprint images with the fingerprint database image of the finger position corresponding to the fingerprint image, wherein the enlarged fingerprint image is obtained by at least one scale enlargement of the fingerprint image
  • the reduced fingerprint image is obtained by scaling down the fingerprint image at least once, and the 2D fingerprint image is obtained by establishing a corresponding 3D fingerprint model for the fingerprint image, according to the 3D fingerprint model corresponding to the fingerprint image. , obtained by expanding the fingerprint image.
  • Item 12 The non-contact fingerprint identification method according to item 11, wherein the method further comprises: a living body identification step;
  • the shooting of the one or more fingers of the object to be recognized includes: shooting one or more fingers of the object to be recognized under the condition of turning off the flash, obtaining a first finger image, and turning on the flash of the object to be recognized one or more fingers of the camera are taken to obtain a second finger image;
  • the living body identification step includes: performing Fourier transform on the first fingerprint image and the second fingerprint image to obtain a frequency domain signal; performing brightness analysis on each area of the first fingerprint image and the second fingerprint image to obtain The change value of the brightness of the corresponding areas of the first fingerprint image and the second fingerprint image when the flashlight is turned off and the flashlight is turned on; the stacked image obtained by stacking the first fingerprint image and the second fingerprint image, the frequency
  • the domain signal and the luminance value are input into a deep neural network, and a first judgment result representing whether the recognized object is a living body is obtained through the deep neural network; a recognition result is obtained according to at least one judgment result, and the at least one judgment result includes the first judgment result;
  • the comparing the fingerprint image with the fingerprint database image includes, when the recognition result shows that the recognized object is a living body, comparing the fingerprint image with the fingerprint database image;
  • the first fingerprint image is obtained by performing image processing on the first finger image
  • the second fingerprint image is obtained by performing image processing on the second finger image.
  • a contactless fingerprint identification method comprising:
  • the fingerprint data includes a fingerprint image and accompanying information
  • Item 14 The non-contact fingerprint identification method according to Item 13, further comprising: if it is detected that the fingerprint image is collected non-contact, performing coarse normalization and fine normalization on the fingerprint image; and/or If it is detected that the fingerprint image is captured by contact, fine normalization is performed on the fingerprint image.
  • Item 15 The contactless fingerprint identification method according to Item 14, wherein the coarse normalization comprises:
  • the size of the fingerprint image after removing the background is adjusted according to the length of the first knuckle line or the area of the foreground area, wherein the first knuckle line is determined by a knuckle line model, so
  • the fingertip outline is formed by acquiring a plurality of boundary points of the edge of the finger through an outline model and connecting the boundary points, and the foreground area is formed by the fingertip outline and the first finger. The area enclosed by nodal lines.
  • Item 16 The non-contact fingerprint identification method according to Item 14 or 15, the fine normalization comprising: detecting the ridge density of a plurality of sub-regions included in the fingerprint image; and
  • the overall frequency is calculated according to the ridge line density of the plurality of sub-regions, and the overall frequency is adjusted to a preset value.
  • Item 17 The non-contact fingerprint identification method according to Item 16, wherein detecting the ridge line density of the plurality of sub-regions included in the fingerprint image comprises:
  • the ridge density of each of the plurality of sub-regions is calculated according to the minimum translation-invariant distance.
  • a device for non-contact fingerprint recognition comprising:
  • an image acquisition module configured to: photograph one or more fingers of the object to be recognized to obtain a finger image including the one or more fingers;
  • an identification module configured to: identify fingertip positions of the one or more fingers in the finger image according to a finger identification model to obtain one or more fingers corresponding to the one or more fingers a fingertip frame comprising the fingertip area from the distal-most end of the one or more fingers to the first knuckle line of the one or more fingers; from the finger images according to the one or more fingertip frames obtain one or more fingertip images corresponding to fingertip positions of the one or more fingers;
  • a processing module configured to: perform image processing on the one or more fingertip images to obtain a fingerprint image
  • the comparison module is configured to: compare the fingerprint image with the fingerprint database image to obtain a comparison result, including:
  • the fingerprint image is sent to the server, so that the server compares the fingerprint image with the fingerprint database image to obtain a comparison result.
  • a terminal comprising:
  • the at least one memory is used for storing program codes
  • the at least one processor is used for calling the program code stored in the at least one memory to execute the method of any one of items 1 to 17.
  • Item 20 A storage medium for storing program code for performing the method of any one of items 1 to 17.
  • the method includes: photographing one or more fingers of the object to be recognized to obtain a finger image including the one or more fingers; identifying the one or more fingers in the finger image according to a finger recognition model fingertip positions of a plurality of fingers to obtain one or more fingertip frames corresponding to the one or more fingers, the fingertip frames including the most distal ends of the one or more fingers to the the fingertip area of the first phalanx line of the one or more fingers; obtain the fingertip position related to the one or more fingers from the finger image according to the one or more fingertip boxes corresponding one or more fingertip images; performing image processing on the one or more fingertip images to obtain a fingerprint image; and comparing the fingerprint image with a fingerprint library image to obtain a comparison result , including: locally comparing the fingerprint image with the fingerprint database image to obtain a comparison result; or sending the fingerprint image to the server, so that the server compares the fingerprint image with the fingerprint database image, to obtain comparison results.
  • the method includes: photographing one or more fingers of the object to be recognized to obtain a
  • the contactless fingerprint identification method, device, terminal and storage medium of the present application are reproducible and can be used in various industrial applications.
  • the non-contact fingerprint identification method, device, terminal and storage medium of the present application can be used in any application that needs to detect fingerprints.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Collating Specific Patterns (AREA)

Abstract

La présente invention concerne un procédé et un appareil de reconnaissance d'empreintes digitales sans contact, un terminal et un support d'informations. Le procédé comprend : la photographie d'un ou de plusieurs doigts d'un sujet à identifier pour acquérir une image de doigt du ou des doigts ; sur la base d'un modèle de reconnaissance de doigt, la reconnaissance de la position de bout de doigt du ou des doigts dans l'image de doigt pour acquérir une ou plusieurs images de bout de doigt correspondant au ou aux doigts, l'image de bout de doigt comprenant une zone d'empreinte digitale de l'extrémité la plus distale du ou des doigts à un premier pli interphalangien du ou des doigts ; sur la base de la ou des images de bout de doigt, l'acquisition à partir de l'image de doigt d'une ou de plusieurs images d'empreinte digitale correspondant à la position de bout de doigt du ou des doigts ; la réalisation d'un traitement d'image sur la ou les images de bout de doigt pour acquérir une image d'empreinte digitale ; et la comparaison de l'image d'empreinte digitale à une image de bibliothèque d'empreintes digitales pour acquérir un résultat de comparaison, comprenant : la comparaison locale de l'image d'empreinte digitale avec une image de bibliothèque d'empreintes digitales pour acquérir un résultat de comparaison ; ou l'envoi de l'image d'empreinte digitale à un serveur, de telle sorte que le serveur compare l'image d'empreinte digitale à une image de bibliothèque d'empreintes digitales pour acquérir un résultat de comparaison. La présente invention peut mettre en œuvre efficacement une reconnaissance d'empreintes digitales sans contact et améliorer l'expérience de l'utilisateur.
PCT/CN2021/122240 2020-09-30 2021-09-30 Procédé et appareil de reconnaissance d'empreintes digitales sans contact, terminal, et support d'informations Ceased WO2022068931A1 (fr)

Applications Claiming Priority (12)

Application Number Priority Date Filing Date Title
CN202011069036.7A CN112232163B (zh) 2020-09-30 2020-09-30 指纹采集方法及装置、指纹比对方法及装置、设备
CN202011061125.7A CN112232157B (zh) 2020-09-30 2020-09-30 指纹区域检测方法、装置、设备、存储介质
CN202011056418.6A CN112232152B (zh) 2020-09-30 2020-09-30 非接触式指纹识别方法、装置、终端和存储介质
CN202011057702.5 2020-09-30
CN202011062601.7A CN112232159B (zh) 2020-09-30 2020-09-30 指纹识别的方法、装置、终端及存储介质
CN202011056390.6A CN112016525A (zh) 2020-09-30 2020-09-30 非接触式指纹采集方法和装置
CN202011061125.7 2020-09-30
CN202011056390.6 2020-09-30
CN202011056418.6 2020-09-30
CN202011057702.5A CN112232155B (zh) 2020-09-30 2020-09-30 非接触指纹识别的方法、装置、终端及存储介质
CN202011069036.7 2020-09-30
CN202011062601.7 2020-09-30

Publications (1)

Publication Number Publication Date
WO2022068931A1 true WO2022068931A1 (fr) 2022-04-07

Family

ID=80951253

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/122240 Ceased WO2022068931A1 (fr) 2020-09-30 2021-09-30 Procédé et appareil de reconnaissance d'empreintes digitales sans contact, terminal, et support d'informations

Country Status (1)

Country Link
WO (1) WO2022068931A1 (fr)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115578760A (zh) * 2022-11-15 2023-01-06 山东圣点世纪科技有限公司 一种基于地形起伏度静脉识别的控制系统及控制方法
CN116229523A (zh) * 2022-12-28 2023-06-06 泉州装备制造研究所 一种指纹姿态估计方法、装置和设备
CN117689587A (zh) * 2022-08-26 2024-03-12 腾讯科技（深圳）有限公司 一种图像处理方法、设备、存储介质及计算机程序产品
CN117827076A (zh) * 2024-03-04 2024-04-05 上海海栎创科技股份有限公司 一种多指并拢触摸区域分割方法
CN118429605A (zh) * 2024-07-04 2024-08-02 国网浙江省电力有限公司永康市供电公司 基于分析模型的设备异常分析方法
CN120510610A (zh) * 2025-05-13 2025-08-19 成都西部泰力智能设备股份有限公司 基于人工智能的食品生产监控图像异常识别方法及系统

Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101504724A (zh) * 2009-03-20 2009-08-12 北京中星微电子有限公司 一种指纹对准的方法、指纹采集装置、指纹对准装置
CN104598870A (zh) * 2014-07-25 2015-05-06 北京智膜科技有限公司 基于智能移动信息设备的指纹活体检测方法
CN105279504A (zh) * 2015-11-30 2016-01-27 南昌欧菲生物识别技术有限公司 指纹识别装置、指纹识别方法及其终端设备
CN107438854A (zh) * 2015-02-06 2017-12-05 维里迪乌姆Ip有限责任公司 使用移动设备捕获的图像执行基于指纹的用户认证的系统和方法
CN108009520A (zh) * 2017-12-21 2018-05-08 东南大学 一种基于卷积变分自编码器神经网络的手指静脉识别方法及系统
CN110008892A (zh) * 2019-03-29 2019-07-12 北京海鑫科金高科技股份有限公司 一种基于四连指指纹图像采集的指纹核验方法及装置
CN110287918A (zh) * 2019-06-28 2019-09-27 Oppo广东移动通信有限公司 活体识别方法及相关产品
CN110728201A (zh) * 2019-09-20 2020-01-24 南京元初科技有限公司 用于指纹识别的图像处理方法及装置
CN111179265A (zh) * 2020-01-10 2020-05-19 敦泰电子（深圳）有限公司 基于图像的指纹质量评估方法、装置及电子设备
CN111582228A (zh) * 2020-05-20 2020-08-25 深圳前海微众银行股份有限公司 活体掌纹的识别方法、装置、设备及存储介质
CN112016525A (zh) * 2020-09-30 2020-12-01 墨奇科技（北京）有限公司 非接触式指纹采集方法和装置
CN112232157A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 指纹区域检测方法、装置、设备、存储介质
CN112232155A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 非接触指纹识别的方法、装置、终端及存储介质
CN112232163A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 指纹采集方法及装置、指纹比对方法及装置、设备
CN112232152A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 非接触式指纹识别方法、装置、终端和存储介质
CN112232159A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 指纹识别的方法、装置、终端及存储介质

Patent Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101504724A (zh) * 2009-03-20 2009-08-12 北京中星微电子有限公司 一种指纹对准的方法、指纹采集装置、指纹对准装置
CN104598870A (zh) * 2014-07-25 2015-05-06 北京智膜科技有限公司 基于智能移动信息设备的指纹活体检测方法
CN107438854A (zh) * 2015-02-06 2017-12-05 维里迪乌姆Ip有限责任公司 使用移动设备捕获的图像执行基于指纹的用户认证的系统和方法
CN105279504A (zh) * 2015-11-30 2016-01-27 南昌欧菲生物识别技术有限公司 指纹识别装置、指纹识别方法及其终端设备
CN108009520A (zh) * 2017-12-21 2018-05-08 东南大学 一种基于卷积变分自编码器神经网络的手指静脉识别方法及系统
CN110008892A (zh) * 2019-03-29 2019-07-12 北京海鑫科金高科技股份有限公司 一种基于四连指指纹图像采集的指纹核验方法及装置
CN110287918A (zh) * 2019-06-28 2019-09-27 Oppo广东移动通信有限公司 活体识别方法及相关产品
CN110728201A (zh) * 2019-09-20 2020-01-24 南京元初科技有限公司 用于指纹识别的图像处理方法及装置
CN111179265A (zh) * 2020-01-10 2020-05-19 敦泰电子（深圳）有限公司 基于图像的指纹质量评估方法、装置及电子设备
CN111582228A (zh) * 2020-05-20 2020-08-25 深圳前海微众银行股份有限公司 活体掌纹的识别方法、装置、设备及存储介质
CN112016525A (zh) * 2020-09-30 2020-12-01 墨奇科技（北京）有限公司 非接触式指纹采集方法和装置
CN112232157A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 指纹区域检测方法、装置、设备、存储介质
CN112232155A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 非接触指纹识别的方法、装置、终端及存储介质
CN112232163A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 指纹采集方法及装置、指纹比对方法及装置、设备
CN112232152A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 非接触式指纹识别方法、装置、终端和存储介质
CN112232159A (zh) * 2020-09-30 2021-01-15 墨奇科技（北京）有限公司 指纹识别的方法、装置、终端及存储介质

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117689587A (zh) * 2022-08-26 2024-03-12 腾讯科技（深圳）有限公司 一种图像处理方法、设备、存储介质及计算机程序产品
CN115578760A (zh) * 2022-11-15 2023-01-06 山东圣点世纪科技有限公司 一种基于地形起伏度静脉识别的控制系统及控制方法
CN116229523A (zh) * 2022-12-28 2023-06-06 泉州装备制造研究所 一种指纹姿态估计方法、装置和设备
CN117827076A (zh) * 2024-03-04 2024-04-05 上海海栎创科技股份有限公司 一种多指并拢触摸区域分割方法
CN118429605A (zh) * 2024-07-04 2024-08-02 国网浙江省电力有限公司永康市供电公司 基于分析模型的设备异常分析方法
CN120510610A (zh) * 2025-05-13 2025-08-19 成都西部泰力智能设备股份有限公司 基于人工智能的食品生产监控图像异常识别方法及系统

Similar Documents

Publication Publication Date Title
CN112232155B (zh) 非接触指纹识别的方法、装置、终端及存储介质
CN112016525A (zh) 非接触式指纹采集方法和装置
US20220215686A1 (en) Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices
JP7242528B2 (ja) モバイルデバイスを用いてキャプチャした画像を使用する、指紋によるユーザ認証を実施するためのシステムおよび方法
CN112232163B (zh) 指纹采集方法及装置、指纹比对方法及装置、设备
KR102587193B1 (ko) 모바일 장치를 사용하여 촬영된 이미지를 이용한 지문-기반 사용자 인증 수행 시스템 및 방법
US10339362B2 (en) Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices
WO2022068931A1 (fr) Procédé et appareil de reconnaissance d'empreintes digitales sans contact, terminal, et support d'informations
CN112232159B (zh) 指纹识别的方法、装置、终端及存储介质
CN110008813B (zh) 基于活体检测技术的人脸识别方法和系统
WO2020258121A1 (fr) Procédé et appareil de reconnaissance faciale et dispositif électronique
CN106529436B (zh) 一种身份一致性认证方法、装置和移动终端
CN107122744A (zh) 一种基于人脸识别的活体检测系统及方法
CN108304789A (zh) 脸部识别方法及装置
CN111598065A (zh) 深度图像获取方法及活体识别方法、设备、电路和介质
CN112232152B (zh) 非接触式指纹识别方法、装置、终端和存储介质
CN119007252A (zh) 一种非接触式的多模态融合的生物识别系统、方法和装置
CN116311400A (zh) 掌纹图像的处理方法、电子设备及存储介质
CN112232157B (zh) 指纹区域检测方法、装置、设备、存储介质
Barra et al. Unconstrained ear processing: What is possible and what must be done
CN212569821U (zh) 非接触式指纹采集装置
Zhang et al. Advanced Palmprint Authentication
Das et al. Liveness Detection in Iris Using CNN
Dixit et al. SIFRS: Spoof Invariant Facial Recognition System (A Helping Hand for Visual Impaired People)
CN120687999A (zh) 一种基于大模型多模态场馆的身份认证方法及系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21874585

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 11/07/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21874585

Country of ref document: EP

Kind code of ref document: A1