JPS6329888A - Feature extracting device for finger print - Google Patents

Abstract

PURPOSE:To extract the feature point of an input finger print by a simple means by forming a peak image corresponding to an uplift line and a trough image corresponding to the part other than the uplift part from an inputted finger print image and extracting respective end points of the trough and peak images as feature points. CONSTITUTION:Image information corresponding to the uplift pattern of a finger print is read out from an image reader 11 and converted into digital image data by an A/D converter 12 and the image data are stored in an image memory 13. The image data stored in the image memory 13 are supplied to a microcomputer 14 for executing the extracting and collating operation of the feature points of the finger print. Registration pattern data obtained from an EEPROM 15 storing and setting the feature points of the registered finger print are supplied to the microcomputer 14 and the microcomputer 14 is provided with a ROM 16 for storing program data or the like to execute the extracting and collating the operation of the feature points and a RAM 17 for storing data or the like to be used in an arithmetic process.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a fingerprint feature extraction device that facilitates fingerprint verification.

The use of fingerprints as a key is being considered as a means of controlling entry. As a fingerprint verification device used for such room entry management, etc., it is essential to extract fingerprint minutiae that represent information about each individual based on a fingerprint image. The end points and branch points of the ridges of the fingerprint are used as the minutiae.

Generally, when performing room entry control, an image reading device is used to read the fingerprint image of the person attempting to enter the room, but it is difficult to read this fingerprint image in a sufficiently clear state. There is usually "bleeding" or "fading" in the input fingerprint image.

The usual processing means for fingerprint input images is to binarize the input image information, further thin the line, and extract end points and branch points of the thin line image (skeletalized image) thus obtained. However, due to the above-mentioned "bleeding" and "blurring", many "pseudo minutiae" other than the original minutiae points due to end points and branch points are obtained.

Therefore, highly accurate fingerprint matching cannot be performed in such a state, and in order to improve the accuracy of fingerprint matching, it is necessary to remove the pseudo minutiae points by some means.

As a method for removing such pseudo minutiae, for example, as shown in Japanese Unexamined Patent Publication No. 59-144982, an image in which the ridge side of a fingerprint is thinned (mountain image) and a thinning in other than the ridges are used. A method has been considered in which feature points are removed using an image (valley image) expressed as a valley image. However, with such means, in each image of a mountain image and a valley image,
It is necessary to detect pixel points at end points and branch points, which requires complicated steps and a lot of processing time.

[Problems to be Solved by the Invention] This invention has been made in view of the above-mentioned points, and aims to extract minutiae of an input fingerprint by a sufficiently simple means. The present invention aims to provide a fingerprint feature extraction device that can be effectively used not only for room entry management, but also as a keyless system for vehicles.

[Means for Solving the Problems] That is, in the fingerprint feature extraction device according to the present invention, mountain images corresponding to fingerprint ridges and portions other than the ridges are extracted from an input fingerprint image. A valley image is created, and the end points of each of the mountain image and valley image are extracted as feature points. In this case, a section specified along the skeleton line is determined from the end points extracted above, and if a new end point or branch point is detected in the specified section, the end point is treated as a pseudo feature point. judge.

Then, feature points consisting of end points of each of the mountain image and the valley image excluding the pseudo feature points are compared with the registered pattern. Here, when this device is used particularly for a vehicle, a switch means operated in response to a fingerprint input operation is provided, and based on a signal from this switch means,
Only during the fingerprint input operation, power is supplied from the vehicle-mounted battery to the image input means and further to the feature point detection operation section.

[Operation] In the fingerprint feature extraction device as described above, the branch points of the mountain image corresponding to the ridges of the fingerprint correspond to the end points of the valley image, and furthermore, the branch points of the valley image correspond to the end points of the valley image. The fingerprint matching operation can be performed using a pattern of end points that can be extracted relatively easily without detecting these branch points. In other words, endpoints that are detected based on pseudo images included in the input image are determined by checking the existence of other endpoints or branching points within the section specified along the skeleton line from the endpoint. , it is possible to determine whether the point is a pseudo feature point or not. Therefore, it is no longer necessary to discriminate pseudo minutiae based on branch points, minutiae patterns can be easily extracted, and highly accurate fingerprint matching can be performed effectively in a short processing time. .

In addition, when this device is used as a means to replace a vehicle key, the vehicle's on-board battery power supply, which has a limited capacity, is effectively used only during the fingerprint verification operation, for example, when the device is installed on a door. It can be effectively used as an unlocking means.

[Embodiments of the invention] Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

This embodiment shows an example in which it is applied to a keyless entry system that allows a car to open a locked door without using a key. [
is installed, for example, at the handle part of the door.

The configuration of this image reading device 11 will be described later, but image information corresponding to the ridge pattern of a fingerprint is read out from this image reading device 11, and this image information is read out by an A/D converter 12. converted to digital image data,
This image data is stored in the image memory 13. Then, the image data stored in the image memory 13 is supplied to the microcomputer 14 which executes the operation of extracting the minutiae of the fingerprint and the operation of matching.

The microcomputer 14 is supplied with registered pattern data from an EEPROM 15 in which minutiae points of registered fingerprints are stored and set, and is further configured to carry out the above-mentioned minutiae extraction and verification operations. A ROM 1B that stores program data and the like, and a RAM 17 that stores data and the like used in the calculation process are provided.

When the microcomputer 14 determines that the input fingerprint data matches the registered data, a signal is outputted via the interface 18, and via a switch circuit 19 such as a relay, Electric power from the battery power source 20 is supplied to an actuator 21 that performs door unlock control, so that the door unlocking operation is performed.

Further, in connection with the image reading device 11 that performs the fingerprint detection operation, a touch sensor 22 is set which is activated when a finger is placed in a predetermined place. The detection signal is converted into a predetermined operation command signal by the detection processing circuit 23. A command signal based on the detection signal from the touch sensor 22 operates a switch circuit 24 composed of a relay or the like, and supplies power from the battery power supply 20 to the constant voltage power supply 25 to obtain a predetermined voltage. Make sure that the power supply voltage is generated. This constant voltage power supply 25 supplies operating power to the fingerprint reading processing section including the microcomputer, as well as the lamp 2B that illuminates the fingertip when reading a fingerprint.

The battery charger [20 is the ignition switch 27
It is supplied to an engine operating circuit, etc., including a starter, etc., which are not shown in this figure. Further, the microcomputer 14 is configured to receive an operation command signal from an operation section 28 for inputting an appropriate operation command, for example, for manually inputting a fingerprint registration command.

FIG. 2 shows a specific example of the configuration of the fingerprint image reading device 11, in which the fingerprint detection surface is constituted by a prism 31. The fingertip 32 on which a fingerprint is to be detected is pressed against one surface of the prism 3I. This prism 31 portion is provided with a cover 33 which serves both to prevent the surface of the prism 31 serving as a detection portion from becoming dirty and to serve as a guide for the finger 32.

The light beam from the lamp 26 is converted into parallel light beams through a lens system 34, a slit 35, etc., and is input to the prism 31 via a mirror 3B.
The second fingerprint surface is illuminated.

Further, the image of the fingerprint surface of the fingertip 32 illuminated in this way is guided to an imaging unit 39 such as a CCD via a mirror 37 and a lens system 38 and is imaged. Image information corresponding to the ridges is output.

Here, the portion of the prism 11 corresponding to the detection surface is contacted when the fingertip 32 is inserted into the cover 33,
A touch sensor 22 is provided to be operated.

In the device configured as described above, fingerprint registration is first performed. Fingertip 3 where the fingerprint you are trying to print exists
Set 3. When the fingertip is set in this manner, the presence of the fingertip is detected by the touch sensor 22, the switch circuit 24 is turned on, and the battery power source 20 is connected to the constant voltage power source 25. That is, lamp 26
is turned on, power is supplied to the part that executes the fingerprint feature detection operation, including the microcomputer 14, and the part is set to be in an operating state. comes to be imaged by the CCD imaging section 39. In this case, the detection processing circuit 23 to which the detection signal from the touch sensor 22 is supplied turns on the switch circuit 24 within a predetermined time range after the detection signal from the touch sensor 22 rises. Within the time range necessary for the fingerprint image to be read, the image reading device 11 and the microcomputer 14
The part is set to an operational state.

Image information based on the image corresponding to the ridges of the fingerprint read by the image reading device 11 is converted into digital image data by the A/D converter 12 and stored in the image memory 13.

If you operate the registration button set on the operation unit 28 in this state, the command from the operation unit 28 will cause the
The microcomputer 14 reads and processes the fingerprint image data stored in the image memory 13, and stores the pattern of minutiae of this fingerprint image in the EEFROM 15. In this EEPROM 15, for example, the fingerprints of a plurality of people who will be using this vehicle are appropriately registered and set.

With the fingerprint registration completed in this way, if the person using the vehicle places his or her fingertip on a predetermined part of the fingerprint image reading device 11, the fingerprint will be read in the same manner as the registration described above and stored in the image memory 13. will be stored in . The microcomputer 14 executes fingerprint processing in the same manner as at the time of registration, and stores the minutiae pattern of the human fingerprint and the EEPROM.
The feature patterns stored in 15 are sequentially compared.

If it is confirmed that a feature pattern matching the feature pattern of the input fingerprint is registered, the microcomputer 14 generates an output to the interface 18, turns on the switch circuit 19, and performs the door unlocking operation. This will enable the execution of the following.

In other words, the user simply lightly touches the reading section of the fingerprint image reading device 11 with his or her finger to unlock the car door, and at this time, it is ensured that the individual is not identified. Since it is identified, it also has the effect of preventing theft.

The minutiae of a fingerprint will be explained below. (A) in Figure 3
1 shows an example of an input fingerprint image, and this image is a binary image obtained by binarizing the fingerprint image input to the image memory using an appropriate threshold value.

From the fingerprint image shown in FIG. 3A, a mountain image as shown in FIG. 3B, which corresponds to the ridges of the fingerprint, can be obtained as skeleton lines. This mountain image is then used as read image data. Furthermore, if the portions that are not ridges are extracted as skeleton lines, a valley image as shown in FIG. 3(C) will be obtained corresponding to the valley portions between the ridges.

This figure 3 shows the case where there is no pseudo image, and in the image of the fingerprint ridge shown in figure (A), the end point indicated by al and the branch point indicated by a2 are used for matching. This is a feature point.

Here, the means for obtaining the mountain and valley images shown in FIGS. 3B and 3C from the image of the fingerprint ridge shown in FIG.
h) Methods such as the method are known and are produced based on the fingerprint image data stored in the image memory 13.

Looking at the mountain images and valley images shown in (B) and (C) in FIG. 3 above, the feature points a1 and a2 are expressed as the end point b1 of the mountain image and the end point C1 of the valley image, respectively. The end points and branch points, which are the feature points of the fingerprint image, are expressed as the end points of the mountain image and the valley image, respectively.

Figures 4 to 8 each show examples of pseudo feature points, with Figure (A) showing a mountain image and Figure (B) showing a valley image corresponding to Figure (A) above. There is. Figure 4 shows the case where a "beard" exists in the mountain image, and Figures 5 and 6 show the case where a small line segment appears as a pseudo image in a part of the image. The figure and FIG. 8 respectively show cases where "crossing" is created as a pseudo image. And, as is clear from these figures, the fourth
Except for the case of the "whisker" in the figure, two end points are located close to each other in all mountain images or valley images. That is, if two end points exist close to each other, these end points may be removed as pseudo feature points. However, if true minutiae exist close to each other, the true minutiae will also be removed.

For this reason, as shown in FIG. 5, for example, by tracing several points along the skeleton line from the end point 5a detected in the mountain image and finding the direction of this end point, from the end point 5a, A mask 41 as shown by a broken line is set as a section specified along
If this exists, it may be determined that this is a pseudo feature point. Note that this mask 41 is a region along the direction, and indicates a region of a specific neighborhood section that extends on both sides centering on the end point 5a.

In addition, in order to remove the "whiskers" as shown in FIG. 4, if you trace the skeleton line of the "whiskers" from the end point 4a in the mountain image in the same way as above, a branching point 4b will be found in the middle. It begins to appear.

Therefore, when this branch point 4b appears, it is possible to determine that this end point is a pseudo feature point and remove it. In addition, when removing the "crossing" shown in FIG. 8, the direction of this end point is determined by tracing along the skeleton line from the end point 8a detected in the valley image. Then, the mask 4 is moved along the direction from the end point 8a.
1, and when it is determined that another end point 8b exists within this mask 41, it may be determined as a pseudo feature point.

In FIGS. 6 and 7, Ba and 7a indicate detected end points, respectively, and 6b and 7b indicate other end points existing within the mask 41, respectively.

From the end points (4a, 5a, 6as 7a, 8a) detected in either the mountain image or the valley image in this way,
Other endpoints (5b,
6b.

7b, 8b) or the branch point 4b, this detected end point can be determined as a pseudo feature point. At this time, branch points (5a, 5d, 5d,
Be, 6d, 78% 7d, 8c, 8d), it is not necessary to perform equivalence judgment processing, and therefore it is necessary to detect the coordinates of the corresponding branching point and to create pseudo feature points based on this branching point. This eliminates the need for judgment and effectively improves the processing speed of this judgment.

Here, as a means for detecting the end points from the mountain image and the spring image as described above, as illustrated in FIGS. This can be detected by preparing a large number of end point patterns in which the direction of arrangement of "1" is set and sequentially applying such a mask to the entire screen.

Figure 10 shows how to remove pseudo feature points as described above.
This shows the flow of processing in the microcomputer 14. First, in step 101, the end points of the mountain image are detected, and then in step 102, the end points of the spring image are detected.

Then, in the next step 103, the end point coordinates of the mountain image are detected.

Once the coordinates of the end points of the mountain image have been obtained in this way, the next step 200 is to process the pseudo end points of the mountain image.
Then, several points are traced along the skeleton line from the first end point detected above. As a result of tracing in this way, whether or not a branch point exists on the traced skeleton line is monitored in the next step 105. If the existence of a branch point is determined in step 105, In this case, the process proceeds to step 10B, in which the first end point being processed is recognized as a pseudo minutiae. If it is determined in step 105 that there is no branch point, then in step 107 the end point and the direction from the end point after tracing are calculated.

In step 108, a predetermined mask is set in the above direction, and it is determined whether or not other end points exist within this mask. If it is determined that another endpoint exists, the process proceeds to step 106, where the first endpoint being processed is recognized as a pseudo feature point. Also, step 1
If it is determined in step 08 that the end point does not exist within the predetermined mask, the process advances to step 109. In this step 109, it is determined whether or not all end point processing existing in the mountain image has been completed, and if it has not been completed yet, step 103
, the next end point coordinates are detected, and the above process 200 is repeatedly executed.

When the processing for all the endpoints existing in the mountain image is completed in this way, the process proceeds from step 109 to step 110, and the coordinates of the endpoints of the spring image are detected again. If the coordinates of the end points of this spring image are detected, the process proceeds to step 201. In this processing step 201, the same processing as the mountain image processing step 200 is executed, and the detailed flow is omitted here. That is, in this processing step 201, the above processing step 2
Similarly to 00, it is determined in step 110 whether or not the end points of the spring image detected are pseudo feature points, and the process proceeds to step 111 to repeatedly execute the processing of each end point. Then, when it is determined in step 111 that the processing for all end points of this spring image has been completed, the processing for removing the pseudo feature points is completed.

In this way, the coordinates of each feature point by the end points of the mountain image and spring image, excluding pseudo feature points, can be obtained, and by comparing this feature point pattern with the registered feature point pattern, It's good.

[Effects of the Invention] As described above, according to the fingerprint feature extraction device according to the present invention, based on the input fingerprint image, the end points and branch points of the ridges that are the feature points of the input fingerprint image are The end points of the mountain image and the spring image created based on the input fingerprint image are detected, respectively, and the end points that become pseudo feature points can be effectively removed. Therefore, there is no need for means for detecting branch points and for removing pseudo feature points corresponding to branch points, which makes it possible to simplify and effectively shorten processing time. In other words, by relatively simple end point detection, it becomes possible to obtain a true minutiae pattern excluding pseudo minutiae points, making it easy to perform highly accurate fingerprint matching.For example, keyless entry It will be possible to effectively apply it to systems, etc.

In particular, when used for automobiles as shown in the embodiment, it is possible to effectively utilize the limited capacity in-vehicle battery without wasting it, and door unlock control using fingerprints is possible. etc. will be executed reliably and reliably.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram illustrating a fingerprint feature extraction device according to an embodiment of the present invention, FIG. 3A to 3C are diagrams showing examples of fingerprint states without pseudo minutiae, white images, and valley images, respectively, and FIGS. 4 to 8 are examples of states with pseudo minutiae, respectively. (A) is a white image (
B) is a diagram showing a valley image, FIGS. 9(A) to (C) are diagrams each showing examples of patterns for detecting end points, and FIG.
FIG. 0 is a flowchart illustrating the flow of pseudo feature point removal processing. 11... Fingerprint image reading device, 13... Image memory, 14... Microcomputer, 15... EPR
OM (registered pattern memory), 19.24...Switch circuit (relay), 20...Battery (for vehicle use),
21...Door unlock control actuator, 22
...Touch sensor, 27...Ignition switch. Applicant's representative Patent attorney Takehiko Suzue Figure 1 Figure 2

Claims (3)

[Claims] (1) an image input device for reading a fingerprint image; a means for detecting end points of skeleton lines of the ridge image based on a crest image corresponding to the ridges of the fingerprint input from the input device; means for detecting the end points of the skeleton line of the valley image corresponding to the valley portion between the ridges, which is created based on the image of the ridges; means for determining the detected end point as a pseudo minutiae in a state where the existence of another endpoint or branch point is confirmed within the section specified along the line, and determining the detected end point as a pseudo minutiae point; A fingerprint feature extraction device characterized in that the end points of each of the mountain image and the valley image except for the end points of the mountain image and the valley image are taken as feature points, and this feature point pattern is compared with a registered pattern. (2) an image input device for reading a fingerprint image; a means for detecting end points of the ridge image based on a ridge image corresponding to the ridge of the fingerprint input from the input device; Means for detecting an end point of a valley image corresponding to a valley portion between the ridges, which is created based on a line image; A means for determining the detected end point as a pseudo feature point in a state where the existence of other end points or branch points is confirmed in the section, and a pattern of end points of each of the mountain image and valley image excluding the pseudo feature point. a means for comparing the pattern with the registered pattern; an in-vehicle control mechanism that is powered by in-vehicle battery power upon detection of a match by the comparison means; and an in-vehicle control mechanism that is activated in response to a fingerprint input operation. 1. A fingerprint feature extraction device comprising: switch means configured to supply operating power from the battery to the image input device during a fingerprint input operation. (3) The fingerprint feature extraction device according to claim 2, wherein the image input device is set at a door of a vehicle, and the in-vehicle control mechanism is constituted by an unlocking mechanism for the door.