WO2019024771A1 - 车险图像处理方法、装置、服务器及系统 - Google Patents

车险图像处理方法、装置、服务器及系统 Download PDF

Info

Publication number
WO2019024771A1
WO2019024771A1 PCT/CN2018/097336 CN2018097336W WO2019024771A1 WO 2019024771 A1 WO2019024771 A1 WO 2019024771A1 CN 2018097336 W CN2018097336 W CN 2018097336W WO 2019024771 A1 WO2019024771 A1 WO 2019024771A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
car insurance
classification
vehicle
image processing
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/CN2018/097336
Other languages
English (en)
French (fr)
Inventor
侯金龙
章海涛
郭昕
徐娟
王剑
程远
程丹妮
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.)
Alibaba Group Holding Ltd
Original Assignee
Alibaba Group Holding 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
Application filed by Alibaba Group Holding Ltd filed Critical Alibaba Group Holding Ltd
Priority to MYPI2019005049A priority Critical patent/MY202176A/en
Priority to EP18840973.4A priority patent/EP3621029A4/en
Priority to KR1020197027479A priority patent/KR102373251B1/ko
Priority to JP2019550188A priority patent/JP7118082B2/ja
Priority to SG11201908245Q priority patent/SG11201908245QA/en
Publication of WO2019024771A1 publication Critical patent/WO2019024771A1/zh
Priority to PH12019502055A priority patent/PH12019502055A1/en
Priority to US16/719,329 priority patent/US10846556B2/en
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
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51—Indexing; Data structures therefor; Storage structures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/285—Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/10—Image acquisition
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/19—Recognition using electronic means
    • G06V30/191—Design or setup of recognition systems or techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
    • G06V30/19173—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40—Document-oriented image-based pattern recognition

Definitions

  • the embodiments of the present specification belong to the field of image data processing technologies, and in particular, to a car insurance image processing method, device, server and system.
  • the insurance service party when the insurance vehicle insurance is out of danger, the insurance service party usually sends the operator or the partner's personnel to the scene of the vehicle to take pictures of the vehicle, the scene of the accident, etc., for subsequent vehicle damage, information verification, etc. .
  • the usual practice includes manually checking the car insurance images and determining the use of each image, such as car damage photos, material damage photos, and ID photos.
  • the embodiment of the present specification aims to provide a car insurance image processing method, device, server and system, which can automatically recognize the scene use of the car insurance image, and quickly and accurately determine the scene classification of the car insurance image.
  • a car insurance image processing method, device, server and system provided by embodiments of the present specification are implemented as follows:
  • a car insurance image processing method comprising:
  • the car insurance image is stored in a classification of the use scene according to the classification label.
  • a car insurance image processing device comprising:
  • An image acquisition module configured to acquire a car insurance image
  • An image processing module configured to process the vehicle risk image by using a preset image classification algorithm to determine at least one category label of the vehicle risk image
  • a classification storage module configured to store the car insurance image in a classification of the usage scenario according to the classification label.
  • a car insurance image processing apparatus includes a processor and a memory for storing processor executable instructions, the processor implementing the instructions to:
  • the car insurance image is stored in a classification of the use scene according to the classification label.
  • a server comprising at least one processor and a memory for storing processor-executable instructions, the processor implementing the instructions to:
  • the car insurance image is stored in a classification of the use scene according to the classification label.
  • a car insurance image system includes a captured image storage unit, an algorithm server, and a car insurance image database, the algorithm server including at least one processor and a memory for storing processor executable instructions, the processor implementing the instructions :
  • a car insurance image processing method, device, server and system provided by one or more embodiments of the present specification can automatically process a car insurance image by using a selected image classification algorithm to identify a classification to which the car insurance image belongs.
  • the accuracy of the classification of the car insurance image and the labeling efficiency of the car insurance image can be greatly improved, the time for the manual recognition process is reduced, and the accuracy and reliability of the image processing of the car insurance are higher.
  • FIG. 1 is a schematic diagram of a system framework of an application vehicle image processing method provided by the present specification
  • FIG. 2 is a schematic flow chart of an embodiment of a car insurance image processing method provided by the present specification
  • FIG. 3 is a schematic diagram of image processing using a single task deep convolutional neural network classification model in one embodiment of the present specification
  • FIG. 4 is a schematic structural diagram of a multi-task deep convolutional neural network classification model of a shared convolution layer provided in one embodiment of the method described in the present specification;
  • FIG. 5 is a schematic flowchart diagram of an embodiment of a service processing method provided by the present specification.
  • FIG. 6 is a schematic diagram of an implementation scenario of an embodiment of the method in the present specification.
  • FIG. 7 is a block diagram showing the structure of an embodiment of a car insurance image processing apparatus provided by the present specification.
  • FIG. 8 is a block diagram showing the structure of another embodiment of a car insurance image processing apparatus provided by the present specification.
  • FIG. 9 is a schematic structural diagram of a module of another embodiment of a car insurance image processing apparatus provided by the present specification.
  • FIG. 10 is a block diagram showing the structure of an embodiment of a car insurance image server provided by the present specification.
  • the present specification provides method operation steps or device structures as shown in the following embodiments or figures, there may be more or partial merged fewer operational steps in the method or device based on conventional or no inventive labor. Or module unit.
  • the execution order of the steps or the module structure of the device is not limited to the execution order or the module structure shown in the embodiment or the drawings.
  • the device, server or terminal product of the method or module structure is applied, it may be executed sequentially or in parallel according to the method or module structure shown in the embodiment or the drawing (for example, parallel processor or multi-thread processing). Environment, even including distributed processing, server cluster implementation environment).
  • Figure 1 is a system for applying the car insurance image processing method provided in the present specification.
  • the schematic diagram of the framework may include an auto insurance claim image system that stores the operator's on-site collection or obtained from other system databases, or a vehicle insurance image provided by a third party, an algorithm server that identifies and classifies the auto insurance image, and a car insurance claim that stores the auto insurance image classification.
  • Image database can store car insurance image data of a plurality of car insurance businesses, and each car insurance business can include multiple car insurance images.
  • Vehicle insurance images belonging to the same auto insurance business case can be distinguished by a certain same label.
  • the auto insurance claims image system may face the scene of storing a large number of auto insurance images.
  • the auto insurance images collected by these scenes can be divided into different according to the auto insurance case processing needs. Process the scene (or use the scene). Then, the auto insurance image classified by the algorithm can be stored in the set auto insurance claim image database for reading and use in subsequent vehicle damage, nuclear damage and the like.
  • the image may include a general term for various graphics and images, and generally refers to a visually effective picture, which may generally include a paper medium, a film or a photo, a television, and a projection.
  • the image on the instrument or the computer screen or the like, the car insurance image described in this embodiment may include computer image data stored on a readable storage medium after being photographed by a camera or an image capturing device, and may include a vector image, a bitmap, a static, a dynamic Various types of computer images such as images.
  • the automobile insurance image collected by the operator of the vehicle insurance service provider may be uniformly stored in the automobile insurance claims image system as shown in FIG. 1 , and the algorithm server may obtain a part of the automobile insurance business single insurance case or All car insurance images. Then, the car insurance image can be outputted by the preset image classification algorithm, and the attribute information of the plurality of dimensions of the car insurance image can be output.
  • the method may include a scene classification actually used according to the vehicle risk image, such as a vehicle damage image, an object loss image, a voucher, etc., and may also include other key attribute information of the set image, such as a vehicle type, a vehicle body color, a shooting illumination condition, and the like.
  • a scene classification actually used according to the vehicle risk image such as a vehicle damage image, an object loss image, a voucher, etc.
  • other key attribute information of the set image such as a vehicle type, a vehicle body color, a shooting illumination condition, and the like.
  • the method may include:
  • S2 processing the vehicle risk image by using a preset image classification algorithm to determine at least one category label of the vehicle risk image
  • S4 The car insurance image is classified and stored in a usage scenario according to the classification label.
  • the algorithm server for identifying and classifying the car insurance image may acquire the original car insurance image collected from the scene.
  • the manner of obtaining may include obtaining from a unified database, such as the “auto insurance claim image system” described above, or may include an automobile insurance image uploaded by an operator immediately, or obtaining a car insurance image from another server or other third party service party. Implementation scenario.
  • the acquired car insurance image may include a plurality of image format types, a shooting angle, and image information of a plurality of image contents, and may include dozens or even hundreds of car insurance images for a single car insurance business. For example, it may include a panoramic photograph of a plurality of vehicles, a photograph of a part of the damaged part, a detailed photograph, a road and traffic conditions of the surrounding vehicle, a lighting condition, a photo of the person involved in the accident, and the like.
  • the vehicle risk image can be divided into different usage scenarios according to the processing requirements of the automobile insurance business, such as a fixed loss scene, a vehicle model scene, a color scene, and the like.
  • a usage scenario can include multiple category tags. Specifically, for example, in a fixed-loss scenario, a panoramic photo, a part photo, a detailed photo, a frame number, an ID card, a driver's license, a driving license, a scene photograph, etc., may be defined for use in determining damage/nuclear loss. Category; in the vehicle scene, you can define different vehicle types such as SUV, car, bus, truck, etc. In the color scene, you can define different vehicle colors such as black, red, and white. In an implementation scenario, an image can have three categories of labels at the same time: panoramic photo, SUV, black. Certainly, the specific types and numbers of usage scenarios and the specific different categories in the usage scenarios may be determined according to actual image processing requirements and application scenarios.
  • the algorithm server may process the acquired car insurance image, and may use the preset image classification algorithm to identify, classify, and output one or more category tags of the car insurance image.
  • the image classification algorithm may adopt various implementation methods, and may adopt DNN (Deep Nerual Network) or a method based on traditional image features.
  • a deep convolutional neural network model can be used to implement classification of field risk images.
  • the deep convolutional neural network may include a convolutional layer, a pooling layer, an activation function, a fully connected layer, and the like.
  • Other implementations can use mature CNN network models such as Inception-ResNet or custom CNN network models.
  • the algorithm server may first perform image preprocessing on each picture in the case, such as de-averaging, normalization, cropping, etc., and may first remove some obviously non-compliant requirements. Car insurance image.
  • each layer structure of the deep neural network, the convolution kernel size, the backhaul parameters, and the like can be pre-built.
  • the marking data of the vehicle image may be marked with the usage scene category to which the image belongs, the specific classification in the usage scene, the vehicle attribute information, and the shooting conditions/environment.
  • the training pictures used in deep neural network training can be obtained by manual marking of real auto insurance images.
  • the preset image classification algorithm may be a single task deep convolutional neural network classification model, for example, a deep convolutional neural network is used to output a classification result of a category, such as Vehicle color.
  • a deep convolutional neural network NS_1 for identifying a vehicle type a neural convolutional neural network NS_2 for recognizing a vehicle color and a light condition, and other settings may be provided.
  • a deep convolutional neural network NS_3 that identifies the document category.
  • a deep convolutional neural network can be viewed as a single-tasking network model.
  • the preset image classification algorithm may include:
  • S002 A multi-task deep convolutional neural network classification model using a shared convolutional layer.
  • FIG. 4 is a schematic structural diagram of a multi-task deep convolutional neural network classification model of a shared convolutional layer provided in one embodiment of the method described in the present specification.
  • multiple deep convolutional neural network classification models can share convolutional layer parameters and most functional layer parameters (shared parameters are set as needed), and the last few layers of different task classification models (including different The classification of attribute dimensions, if some models are classified by model, and some models are classified by color) parameters may not be shared.
  • the multi-task deep convolutional neural network classification model of the shared convolution layer can automatically classify the massive images in the auto insurance claim scene and automatically acquire the multi-dimensional image attributes. The calculation time of the prediction is greatly reduced, and the image processing speed of the automobile insurance is improved.
  • the classification result output by the algorithm server can be written into the corresponding database to perform classified storage of the usage scenario.
  • corresponding picture P1 is marked with three category labels: panoramic photo, SUV, black, then the picture P1 can be stored in three usage scenarios (fixed loss scene, vehicle scene, color scene), respectively.
  • the picture P1 can be stored in the panoramic photo classification of the timed scene, and can also be stored in the SUV photo classification of the car accident scene, and can also be stored in the black photo classification in the color scene.
  • a relational database including the basic functions of a conventional relational database may be employed to store the classification result of the car insurance image.
  • the basic functions of the relational database may include SELECT (selection), INSERT (insert), ALTER (modification) and other data joint filtering, operations and the like.
  • This storage method allows the operator to select the desired type of image flexibly, quickly and conveniently according to the needs. Therefore, in another embodiment of the method, the classifying storage may include:
  • the image may be filtered according to the actual business scenario. For example, if the fixed loss service needs to be processed at present, an image of three categories of labels of panorama, component, and detail may be set in the relational database. In this way, all the models in each single case can be quickly categorized into panoramas, parts, and detailed photos into the image loss or core loss link, and a large number of interference pictures are automatically filtered out, so that the fixed loss photo collection is cleaner and the processing efficiency is higher. .
  • the relational database it is also possible to manually add auto insurance images and their labels, manually modify the category labels of the photos, and retrieve functions as needed.
  • the method may further include:
  • the text information is stored in association with the car insurance image.
  • FIG. 5 is a schematic flow chart of another method of the method described in the present specification.
  • the embodiment of the present embodiment further introduces a character recognition model (OCR, Optical Character Recognition (also referred to as a character recognition model), and key text information (for example, name, ID number, and address) of the car insurance image. Etc.) for detection, location and identification.
  • OCR character recognition model
  • the corresponding OCR character recognition model can be imported into the photo frame labeled as frame number, ID card, driver's license, driving license, bank card, etc., and the text detection, positioning and recognition are performed, and the obtained result is written into the database.
  • Perform associative storage The associated storage may include information that may be retrieved in the database by the textual information, or other associated information.
  • the ID card number can be used to search for all the car insurance images related to the case of the insured person of the ID card number.
  • the method may further include:
  • the ID card information, the driver's license and the driving license information, the bank card information, etc. can be obtained from the auto insurance image, wherein the ID card information can be used to supervise the anti-money laundering process, and the driving license and the driving license information can be used to confirm the eligibility for payment.
  • Bank card information can be used to ensure that money is properly transferred to the corresponding account.
  • the subsequent process can be automatically entered, so that the image in different usage scenarios can be automatically detected by using the solution of the embodiment, and not only the text information in the photo of the document but also the text information in the photo information can be detected.
  • the key information of the preset type is complete, and the detection result of the key information is recorded.
  • These test results can be fed back to the operators of the auto insurance business. For example, when the key information of the ID card is not detected or the number of digits is not enough, it can be displayed to the operator, so that the operator can quickly locate the missing key information and greatly improve the operator.
  • the auto insurance business handles the efficiency and improves the user experience.
  • FIG. 6 is a schematic diagram of an implementation scenario of an embodiment of the method in the present specification.
  • the image of the car insurance claim acquired on the spot can be divided into three types: a fixed loss photo, a physical damage photo, and a non-determined photo, each type. There can be multiple categories. According to different usage scenarios, such as loss/nuclear damage, auto insurance classification, document photo integrity detection, document number identification, etc., the auto insurance images classified by these tags can be stored in the corresponding usage scenario types. As described in the foregoing embodiments, in some embodiments, a car insurance image may have multiple classification tags that can be stored in different usage scenarios.
  • a car insurance image processing method provided by one or more embodiments of the present specification can automatically process a car insurance image by using a selected image classification algorithm to identify a classification to which the car insurance image belongs.
  • one or more embodiments of the present specification further provide a car insurance image processing device.
  • the apparatus may include a system (including a distributed system), software (applications), modules, components, servers, clients, etc., using the methods described in the embodiments of the present specification, in conjunction with necessary implementation hardware.
  • the apparatus in one or more embodiments provided by the embodiments of the present specification is as described in the following embodiments.
  • the term "unit” or "module” may implement a combination of software and/or hardware of a predetermined function.
  • FIG. 7 is a schematic structural diagram of a module of an embodiment of a vehicle image processing apparatus provided by the present specification. As shown in FIG. 7, the apparatus may include:
  • the image acquisition module 101 can be configured to acquire a car insurance image
  • the image processing module 102 may be configured to process the vehicle risk image by using a preset image classification algorithm to determine at least one category label of the vehicle risk image;
  • the classification storage module 103 can be configured to store the auto insurance image in a classified manner of the usage scenario according to the classification label.
  • the preset image classification algorithm may adopt various methods, such as a classification algorithm based on image features.
  • the image processing module 102 may include:
  • a multi-task classification model module is used for the image classification algorithm using a multi-task deep convolutional neural network classification model of a shared convolutional layer.
  • the classification storage module uses a relational database to store classification results of the automobile risk image.
  • FIG. 8 is a block diagram showing the structure of an embodiment of a car insurance image processing apparatus provided by the present specification.
  • the device may further include:
  • the text recognition module 104 is configured to detect the car insurance image by using a selected optical character recognition algorithm, and identify text information in the car insurance image;
  • the text information storage module 105 can be configured to store the text information in association with the car insurance image.
  • FIG. 9 is a block diagram showing the structure of an embodiment of a car insurance image processing apparatus provided by the present specification. As shown in FIG. 9, the apparatus may further include:
  • the text information detecting module 106 may be configured to detect whether the key information of the preset type in the text information is complete, and record the detection result of the key information.
  • a car insurance image processing device provided by one or more embodiments of the present specification can automatically process a car insurance image by using a selected image classification algorithm to identify a classification to which the car insurance image belongs.
  • the accuracy of the classification of the car insurance image and the labeling efficiency of the car insurance image can be greatly improved, the time for the manual recognition process is reduced, and the accuracy and reliability of the image processing of the car insurance are higher.
  • the above-mentioned user car insurance image processing method or apparatus provided by the embodiments of the present specification may be implemented by a processor executing a corresponding program instruction in a computer, such as a C++ language using a Windows operating system on a server side, a Linux system based server, or the like, for example.
  • a processor executing a corresponding program instruction in a computer
  • a computer such as a C++ language using a Windows operating system on a server side, a Linux system based server, or the like, for example.
  • Another embodiment of a car insurance image processing apparatus provided by the present specification may include a processor and a memory for storing processor-executable instructions, when the processor executes the instructions:
  • the car insurance image is stored in a classification of the use scene according to the classification label.
  • the image classification algorithm executed when the processor executes the instruction may include:
  • a multi-task deep convolutional neural network classification model using shared convolutional layers is described.
  • the car insurance image processing device described in the above embodiment can automatically process the car insurance image by using the selected image classification algorithm to identify the classification to which the car insurance image belongs.
  • the accuracy of the classification of the car insurance image and the labeling efficiency of the car insurance image can be greatly improved, the time for the manual recognition process is reduced, and the accuracy and reliability of the image processing of the car insurance are higher.
  • the method or device described above can be used in a server for processing various car insurance image data, which can greatly improve the accuracy and labeling efficiency of car insurance image classification, and expand the attribute dimension of image classification.
  • a server as shown in FIG. 10, that may include at least one processor and a memory for storing processor-executable instructions that are implemented when the processor executes the instructions:
  • the car insurance image is stored in a classification of the use scene according to the classification label.
  • An embodiment of the system may include acquiring an image storage unit, an algorithm server, and a car insurance image database, the algorithm server including at least one processor and a memory for storing processor executable instructions, the processor executing the The instructions can be implemented:
  • the device or the server or the system described in the foregoing description may further include other embodiments according to the description of the related method embodiments, and the specific implementation manners may refer to the description of the method embodiments, which are not described herein.
  • the various embodiments in the specification are described in a progressive manner, and the same or similar parts between the various embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments.
  • the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment.
  • a car insurance image processing method, apparatus, server and system provided by one or more embodiments of the present specification can automatically process a car insurance image by using a selected image classification algorithm to identify a classification to which the car insurance image belongs.
  • the accuracy of the classification of the car insurance image and the labeling efficiency of the car insurance image can be greatly improved, the time for the manual recognition process is reduced, and the accuracy and reliability of the image processing of the car insurance are higher.
  • PLD Programmable Logic Device
  • FPGA Field Programmable Gate Array
  • HDL Hardware Description Language
  • the controller can be implemented in any suitable manner, for example, the controller can take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (eg, software or firmware) executable by the (micro)processor.
  • computer readable program code eg, software or firmware
  • examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, The Microchip PIC18F26K20 and the Silicone Labs C8051F320, the memory controller can also be implemented as part of the memory's control logic.
  • the controller can be logically programmed by means of logic gates, switches, ASICs, programmable logic controllers, and embedding.
  • Such a controller can therefore be considered a hardware component, and the means for implementing various functions included therein can also be considered as a structure within the hardware component.
  • a device for implementing various functions can be considered as a software module that can be both a method of implementation and a structure within a hardware component.
  • the system, device, module or unit illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product having a certain function.
  • a typical implementation device is a computer.
  • the computer can be, for example, a personal computer, a laptop computer, a car-mounted human-machine interaction device, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet.
  • each module may be implemented in the same software or software and/or hardware when implementing one or more of the specification, or the modules implementing the same function may be implemented by a plurality of sub-modules or a combination of sub-units, etc. .
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or integrated. Go to another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.
  • the controller can be logically programmed by means of logic gates, switches, ASICs, programmable logic controllers, and embedding.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
  • RAM random access memory
  • ROM read only memory
  • Memory is an example of a computer readable medium.
  • Computer readable media including both permanent and non-persistent, removable and non-removable media, can be implemented by any method or technology.
  • the information can be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disk read only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, Magnetic cassette tape, magnetic tape storage, graphene storage or other magnetic storage devices or any other non-transportable media can be used to store information that can be accessed by a computing device.
  • computer readable media does not include temporary storage of computer readable media, such as modulated data signals and carrier waves.
  • one or more embodiments of the present specification can be provided as a method, system, or computer program product.
  • one or more embodiments of the present specification can take the form of an entirely hardware embodiment, an entirely software embodiment or a combination of software and hardware.
  • one or more embodiments of the present specification can employ a computer program embodied on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer usable program code embodied therein. The form of the product.
  • One or more embodiments of the present specification can be described in the general context of computer-executable instructions executed by a computer, such as a program module.
  • program modules include routines, programs, objects, components, data structures, and the like that perform particular tasks or implement particular abstract data types.
  • One or more embodiments of the present specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are connected through a communication network.
  • program modules can be located in both local and remote computer storage media including storage devices.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Business, Economics & Management (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Strategic Management (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Tourism & Hospitality (AREA)
  • Databases & Information Systems (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Medical Informatics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Primary Health Care (AREA)
  • Development Economics (AREA)
  • Technology Law (AREA)

Abstract

本说明书实施例公开了一种车险图像处理方法、装置、服务器及系统。所述车险图像处理方法可以包括：获取事故现场拍摄采集的车险图像，利用预置的图像分类算法对所述车险图像进行识别、分类处理，确定所述车险图像的至少一个类别标签；按照所述分类标签将所述车险图像进行使用场景的分类存储。

Description

车险图像处理方法、装置、服务器及系统 技术领域
本说明书实施方案属于图像数据处理技术领域，尤其涉及一种车险图像处理方法、装置、服务器及系统。
背景技术
随着全国各个城市的快速发展，汽车市场越来越火热，与此同时带动了车险市场的快速发展，车险业务呈明显增加趋势。如何快速、准确、高效的处理车险业务、响应用户需求，是各大保险服务方抢占车险市场的重要保障环节。
目前，在车险理赔业务中，当承保车险出险时，通常保险服务方会派作业人员或委托合作方的人员到出险现场拍摄车辆图像、事故现场图像等，以用于后续车辆定损、信息核实等。目前，车险服务方为了有效管理车辆出险拍摄的车险图像，通常的做法包括采用人工方式审核车险图像，确定出各个图像的用途，例如车损照片、物损照片、证件照片等。随着车险业务的快速增加，服务方需要处理的车险案件越来越多，现场获取的车险图像数据量也越来越多。例如对于一些服务方而言，大约每个车险案例平均拍摄约40张照片，甚至比较复杂的车险案例拍摄数量达到200多张，采用人工进行车险图像识别、分类的处理消耗的人力和时间成本也越来越大。因此，业内需要一种更加快速、准确对车险图像进行处理的方式。
发明内容
本说明书实施例目的在于提供一种车险图像处理方法、装置、服务器及系统，可以自动识别出车险图像的场景用途，快速、准确的确定车险图像的场景分类。
本说明书实施例提供的一种车险图像处理方法、装置、服务器及系统是包括如下的方式实现的：
一种车险图像处理方法，所述方法包括：
获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
按照所述分类标签将所述车险图像进行使用场景的分类存储。
一种车险图像处理装置，所述装置包括：
图像获取模块，用于获取车险图像；
图像处理模块，用于利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
分类存储模块，用于按照所述分类标签将所述车险图像进行使用场景的分类存储。
一种车险图像处理装置，包括处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
按照所述分类标签将所述车险图像进行使用场景的分类存储。
一种服务器，包括至少一个处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
按照所述分类标签将所述车险图像进行使用场景的分类存储。
一种车险图像系统，包括采集图像存储单元、算法服务器、车险图像数据库，所述算法服务器包括至少一个处理器以及用于存储处理器可执行指令的存 储器，所述处理器执行所述指令时实现：
从所述采集图像存储单元获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
将所述车险图像案子所述类别标签存储到所述车险图像数据库相应的使用场景的存储区域。
本说明书一个或多个实施例提供的一种车险图像处理方法、装置、服务器及系统，可以通过选取的图像分类算法自动对车险图像进行处理，识别车险图像所属的分类。利用本说明书实施例方案，可以大幅提高车险图像分类的准确性和车险图像标注效率，减少人工识别处理耗时，车险图像处理的准确性和可靠性更高。
附图说明
为了更清楚地说明本说明书实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是说明书中记载的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本说明书提供的一个应用车险图像处理方法的系统框架示意图；
图2是本说明书提供的所述一种车险图像处理方法实施例的流程示意图；
图3是本说明书一个实施例中使用单一任务的深度卷积神经网络分类模型进行图像处理的示意图；
图4是本说明书所述方法一个实施例中提供的共享卷积层的多任务深度卷积神经网络分类模型结构示意图；
图5是本说明书提供的一种业务处理方法实施例的流程示意图；
图6是本说明书所述方法一种实施例的实施场景示意图；
图7是本说明书提供的一种车险图像处理装置实施例的模块结构示意图；
图8是本说明书提供的另一种车险图像处理装置实施例的模块结构示意图；
图9是本说明书提供的另一种车险图像处理装置实施例的模块结构示意图；
图10是本说明书提供的一种车险图像服务器一个实施例的模块结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本说明书中的技术方案，下面将结合本说明书一个或多个实施例中的附图，对本说明书一个或多个实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是说明书一部分实施例，而不是全部的实施例。基于说明书一个或多个实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都应当属于本说明书实施例方案保护的范围。
虽然本说明书提供了如下述实施例或附图所示的方法操作步骤或装置结构，但基于常规或者无需创造性的劳动在所述方法或装置中可以包括更多或者部分合并后更少的操作步骤或模块单元。在逻辑性上不存在必要因果关系的步骤或结构中，这些步骤的执行顺序或装置的模块结构不限于本说明书实施例或附图所示的执行顺序或模块结构。所述的方法或模块结构的在实际中的装置、服务器或终端产品应用时，可以按照实施例或者附图所示的方法或模块结构进行顺序执行或者并行执行(例如并行处理器或者多线程处理的环境、甚至包括分布式处理、服务器集群的实施环境)。
车险理赔业务处理中，事故现场采集的照片可以统一存储在车险服务方的数据库中，例如图1所示的“车险理赔影像系统”，图1是本说明书提供的一个应用车险图像处理方法的系统框架示意图，可以包括存储作业人员现场采集或者从其他系统数据库获取，或者第三方提供的车险图像的车险理赔影像系统，对车险图像进行识别和分类的算法服务器，以及存储车险图像分类后的车险理赔图像数据库。所述的车险理赔影像系统中可以存储多个车险业务的车险图像 数据，每个车险业务可以包括多张车险图像。属于同一个车险业务案件的车险图像可以采用某个相同的标注予以区分。随着车险业务出不断增加，所述车险理赔影像系统中可能会面临存储着海量车险图像的情景，利用本说明书实施例方案可以将这些现场采集来的车险图像按照车险案件处理需求划分为不同的处理场景(或称为使用场景)。然后经过算法分类后的车险图像可以存储到设置的车险理赔图像数据库中，以供后续车辆定损、核损等作业时读取使用等。
在本说明书一个或多个实施例中，所述的图像可以包括各种图形和影像的总称，通常指具有视觉效果的画面，一般可以包括纸介质上的、底片或照片上的、电视、投影仪或计算机屏幕等上的画面，本实施例中所述的车险图像可以包括通过照相或摄像设备拍摄后存储在可读存储介质上的计算机图像数据，可以包括矢量图、位图，静态、动态图像等多种类型的计算机图像。
下面以某个车险业务出险案例为应用场景对实施方案进行清楚的说明。在本实施例方案中，车险服务方的作业人员事故现场采集的车险图片可以统一存储在如图1所示的车险理赔影像系统中，算法服务器可以从中获取某个车险业务单出险案例的部分或者所有车险图像。然后这车险图像经过预置的图像分类算法的计算，可以输出所述车险图像的多个维度的属性信息。具体的可以包括根据车险图像实际用于的场景分类，如车损图像、物损图像、凭证等，也可以包括设置的图像的其他关键属性信息，例如车型、车身颜色、拍摄光照条件等。具体的一个实施例如图2所示，本说明书提供的一种车险图像处理方法的一个实施例中，所述方法可以包括：
S0：获取车险图像；
S2：利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
S4：按照所述分类标签将所述车险图像进行使用场景的分类存储。
本实施例中，对车险图像进行识别和场景分类处理的算法服务器可以获取从现场采集的原始的车险图像。获取的方式可以包括从统一的数据库中获取， 例如上述所述的“车险理赔影像系统”，也可以包括作业人员即时上传的车险图像，或者从其他服务器或其他第三方服务方中获取车险图像的实施场景。
获取的车险图像可以包括多种图像格式类型、拍摄角度、多种图像内容的图像信息，对于单个车险业务而言，可以包括几十张甚至几百张车险图像。例如可以包括多张车辆的全景照片、受损部位的部件照片、细节照片、出险车辆的周边道路及交通情况、光照条件、事故所涉及到人员的证件照片等。在本实施例中，可以根据车险业务处理需求将车险图像划分为不同的使用场景，如定损场景、车型场景、颜色场景等。在不同的使用场景中，可以进一步划分不同的类别，每个类别可以对应相应的类别标签。一个使用场景可以包括多个类别标签。具体的，例如一个示例中，例如在定损场景中，可以定义全景照片、部件照片、细节照片、车架号、身份证、驾驶证、行驶证、现场照等定损/核损时使用的类别；在车型场景中，可以定义SUV、轿车、公交车、卡车等不同的车辆类型；在颜色场景中，可以定义黑色、红色、白色等不同的车辆颜色。则一个实施场景中，一个图像可以同时拥有三个类别标签：全景照片，SUV，黑色。当然，具体的使用场景种类和数量以及使用场景下的具体不同类别，可以根据实际图像处理需求和应用场景等确定。
算法服务器可以对获取的车险图像进行处理，可以利用预置的图像分类算法对所述车险图像进行识别、分类，输出车险图像的一个或多个类别标签。所述的图像分类算法可以采用多种实现方法，可以采用DNN(Deep Nerual Network，深度神经网络)或者基于传统图像特征的方法。本说明书提供的一个实施例中，可以采用深度卷积神经网络模型实现场险图像的分类。所述的深度卷积神经网络可以包括卷积层、池化层、激活函数、全连接层等。其他的实现方式上可以使用Inception-ResNet等成熟的CNN网络模型，也可以使用定制的CNN网络模型。
当然，在其他的实施例中，所述算法服务器也可以先对该案例中每一张图片进行图片预处理，例如去均值、归一化、裁剪等等，可以先剔除一些明显不 符合要求的车险图像。
一个使用深度神经网络的实施方式中，可以预先构建好深度神经网络的各个层结构、卷积核大小、回传参数等。选取的神经网络的参数可以通过使用打标数据进行小批量梯度下降(mini-batch gradient descent)训练得到，比如mini-batch＝32时，同时以32张训练图片作为输入来训练。车辆图像的打标数据可以标注了该图像所属的使用场景类别、使用场景中的具体分类、车辆属性信息以及拍摄条件/环境等。深度神经网络训练使用的训练图片可以通过对真实车险图像进行人工打标获得。
本说明书所述方法的一个实施例中，所述预置的图像分类算法可以为单一任务的深度卷积神经网络分类模型，例如一个深度卷积神经网络用来输出一种类别的分类结果，如车辆颜色。具体的一个示例中可以如图3所示，可以分别设置一个用于识别车型的深度卷积神经网络NS_1、一个用于识别车辆颜色以及光线情况的神经卷积神经网络NS_2，以及还可以设置其他的例如识别证件类别的深度卷积神经网络NS_3。这种实施方式下，一个深度卷积神经网络可以视为一个单一任务的网络模型。本说明书提供的所述方法的另一个实施例中，所述预置的图像分类算法可以包括：
S002：采用共享卷积层的多任务深度卷积神经网络分类模型。
图4是本说明书所述方法一个实施例中提供的共享卷积层的多任务深度卷积神经网络分类模型结构示意图。在本实施例方案中，多个深度卷积神经网络分类模型可以共享卷积层参数及大部分功能层参数(根据需要设置共享参数)，不同任务的分类模型的最后几层模型(包括涉及不同属性维度的分类，如有的模型以车型分类、有的模型以颜色分类)参数可以不共享。这样，相比单独训练每个任务的模型，本实施例采用共享卷积层的多任务深度卷积神经网络分类模型可以对车险理赔场景中的海量图像进行自动分类以及多维度图像属性自动获取，极大的降低了预测的计算时间，提高了车险图像处理速度。
算法服务器输出的分类结果可以写入到相应的数据库中，进行使用场景的 分类存储。例如上述示例中，对应图片P1，同时标注有三个类别标签：全景照片，SUV，黑色，则该图片P1可以分别存储到三个使用场景(定损场景，车型场景，颜色场景)中，具体的，图片P1可以存储到定时场景的全景照片分类中，同时也可以存储到车险场景的SUV照片分类中，同时也可以存储到颜色场景中的黑色照片分类中。
本说明书提供的所述方法的另一个实施例中，可以采用包含了常规关系型数据库基本功能的关系型数据库来存储车险图像的分类结果。所述的关系型数据库的基本功能可以包括SELECT(选取)、INSERT(插入)、ALTER(修改)等数据联合筛选、操作等处理。这样的存储方式，可以使得作业人员根据需求灵活、快速、便捷的选取所需种类的图像。因此，所述方法的另一种实施例中，所述的分类存储可以包括：
S004：采用关系型数据库存储车险图像的分类结果。
可以根据实际业务场景来筛选图像，例如目前需要处理定损业务，则可以在所述关系型数据库中设置提取全景、部件、细节三个类别标签的图像。这样可以快速调取每一单案例中所有模型分类为全景、部件、细节照片进入到图像定损或者核损环节，自动过滤掉了大量干扰图片，使得定损照片集合更干净，处理效率更高。当然，在所述关系型数据库中，也可以根据需要手动添加车险图像及其标签、手动修改照片的类别标签以及检索等功能。
本说明书所述方法的另一种实施例中，还可以包括：
S6：使用选取的光学字符识别算法对所述车险图像进行检测，识别所述车险图像中的文本信息；
将所述文本信息与所述车险图像进行关联存储。
图5是本说明书所述方法另一个实施例的方法流程示意图。如图5所示，本实施例实施方案还引入文字识别模型(OCR，Optical Character Recognition，光学字符识别，也称为文字识别模型)，对车险图像的关键文字信息(例如姓名、证件号码、住址等)进行检测、定位和识别。具体的，可以对于标注为车架号、 身份证、驾驶证、行驶证、银行卡等证件照片，分别导入对应的OCR文字识别模型，进行文字检测、定位和识别，得到的结果写入数据库，进行关联存储。所述的关联存储可以包括可以通过所述文本信息在数据库中检索到相应的车险图像，或者以及其他向关联的信息。例如可以通过身份证号码搜索到该身份证号码的出险人对应的所有案件相关的车险图像。
本说明书所述方法的另一种实施例中，所述方法还可以包括：
S8：检测所述文本信息中的预设类型的关键信息是否完整，以及记录所述关键信息的检测结果。
本实施例中，还可以自动核实车险业务案件中的关键信息是否完整。例如可以从车险图像中获取身份证信息、驾驶证和行驶证信息、银行卡信息等，其中，身份证信息可以用于监管反洗钱处理，驾驶证、行驶证信息可以用于确认有赔付资格，银行卡信息可以用于确保钱款正确转给对应的账户。对于收集完整的案件可以自动进入后续流程，这样，利用本实施例方案可以自动对不同使用场景中的图像进行自动检测，不仅可以识别例如证件照片中文字信息，还可以检测所述文本信息中的预设类型的关键信息是否完整，以及记录所述关键信息的检测结果。这些检测结果可以反馈给车险业务的作业人员，例如当出现身份证的关键信息没有检测到或者位数不够，则可以展示给作业人员，使作业人员快速定位出缺少的关键信息，大大提高作业人员车险业务处理效率，提高用户使用体验。
图6是本说明书所述方法一种实施例的实施场景示意图，图6中，现场采集获取的车险理赔图片可以分为定损照片、物损照片、非定损照片三个类型，每个类型可以有多个分类。根据不同的使用场景分类，如定损/核损、车险分类、单证类照片完整性检测、证件号码识别等，可以将这些标签分类的车险图像的存储入对应的使用场景类型中。如前述实施例所述，一些实施例中，一个车险图像可以拥有多个分类标签，可以存储到不同的使用场景中。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相 似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。具体的可以参照前述相关处理相关实施例的描述，在此不做一一赘述。
上述对本说明书特定实施例进行了描述。其它实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的结果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连续顺序才能实现期望的结果。在某些实施方式中，多任务处理和并行处理也是可以的或者可能是有利的。
本说明书一个或多个实施例提供的一种车险图像处理方法，可以通过选取的图像分类算法自动对车险图像进行处理，识别车险图像所属的分类。利用本说明书实施例方案，可以大幅提高车险图像分类的准确性和车险图像标注效率，减少人工识别处理耗时，车险图像处理的准确性和可靠性更高。
基于上述所述的用户车险图像处理方法，本说明书一个或多个实施例还提供一种车险图像处理装置。所述的装置可以包括使用了本说明书实施例所述方法的系统(包括分布式系统)、软件(应用)、模块、组件、服务器、客户端等并结合必要的实施硬件的装置。基于同一创新构思，本说明书实施例提供的一个或多个实施例中的装置如下面的实施例所述。由于装置解决问题的实现方案与方法相似，因此本说明书实施例具体的装置的实施可以参见前述方法的实施，重复之处不再赘述。以下所使用的，术语“单元”或者“模块”可以实现预定功能的软件和/或硬件的组合。尽管以下实施例所描述的装置较佳地以软件来实现，但是硬件，或者软件和硬件的组合的实现也是可能并被构想的。具体的，图7是本说明书提供的一种车险图像处理装置实施例的模块结构示意图，如图7所示，所述装置可以包括：
图像获取模块101，可以用于获取车险图像；
图像处理模块102，可以用于利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
分类存储模块103，可以用于按照所述分类标签将所述车险图像进行使用场景的分类存储。
如前述方法所述，预置的图像分类算法可以采用多种方式，如基于图像特征的分类算法等。本说明书所述装置的另一个实施例中，所述图像处理模块102可以包括：
多任务分类模型模块，用于采用共享卷积层的多任务深度卷积神经网络分类模型作为所述图像分类算法。
其他的实施方式中，所述分类存储模块采用关系型数据库存储车险图像的分类结果。
图8是本说明书提供的一种车险图像处理装置实施例的模块结构示意图。所述装置的另一个实施例中，所述装置还可以包括：
文本识别模块104，可以用于使用选取的光学字符识别算法对所述车险图像进行检测，识别所述车险图像中的文本信息；
文本信息存储模块105，可以用于将所述文本信息与所述车险图像进行关联存储。
图9是本说明书提供的一种车险图像处理装置实施例的模块结构示意图。如图9所示，所述装置还可以包括：
文本信息检测模块106，可以用于检测所述文本信息中的预设类型的关键信息是否完整，以及记录所述关键信息的检测结果。
需要说明的，上述所述的装置根据方法实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照相关方法实施例的描述，在此不作一一赘述。
本说明书一个或多个实施例提供的一种车险图像处理装置，可以通过选取的图像分类算法自动对车险图像进行处理，识别车险图像所属的分类。利用本说明书实施例方案，可以大幅提高车险图像分类的准确性和车险图像标注效率，减少人工识别处理耗时，车险图像处理的准确性和可靠性更高。
本说明书实施例提供的上述用户车险图像处理方法或装置可以在计算机中由处理器执行相应的程序指令来实现，如使用windows操作系统的c++语言在服务器端、基于Linux系统的服务器，或其他例如使用android、iOS系统程序设计语言在服务器系统终端实现，以及包括基于量子计算机的处理逻辑实现等。本说明书提供的一种车险图像处理装置的另一种实施例中，可以包括处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
按照所述分类标签将所述车险图像进行使用场景的分类存储。
所述装置的另一个实施例中，所述处理器执行所述指令时执行的图像分类算法可以包括：
采用共享卷积层的多任务深度卷积神经网络分类模型。
需要说明的，上述所述的装置根据方法实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照相关方法实施例的描述，在此不作一一赘述。
上述实施例所述的一种车险图像处理装置，可以通过选取的图像分类算法自动对车险图像进行处理，识别车险图像所属的分类。利用本说明书实施例方案，可以大幅提高车险图像分类的准确性和车险图像标注效率，减少人工识别处理耗时，车险图像处理的准确性和可靠性更高。
上述所述的方法或装置可以用于多种车险图像数据处理的服务器中，可以极大提高车险图像分类的准确度和标注效率，并且扩展了图像分类的属性维度。具体的，本说明书提供一种服务器，如图10所示，可以包括至少一个处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的 至少一个类别标签；
按照所述分类标签将所述车险图像进行使用场景的分类存储。
本说明书还提供一种车险图像系统。所述系统的一个实施例中，可以包括采集图像存储单元、算法服务器、车险图像数据库，所述算法服务器包括至少一个处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时可以实现：
从所述采集图像存储单元获取车险图像；
利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
将所述车险图像案子所述类别标签存储到所述车险图像数据库相应的使用场景的存储区域。
需要说明的是说明书上述所述的装置或服务器或系统根据相关方法实施例的描述还可以包括其他的实施方式，具体的实现方式可以参照方法实施例的描述，在此不作一一赘述。本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于硬件+程序类实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
上述对本说明书特定实施例进行了描述。其它实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的结果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连续顺序才能实现期望的结果。在某些实施方式中，多任务处理和并行处理也是可以的或者可能是有利的。本说明书附图中的虚线部分可以表示在其他的方法或装置实施例中，还可以增加的处理步骤或模块单元。
本说明书一个或多个实施例提供的一种车险图像处理方法、装置、服务器及系统，可以通过选取的图像分类算法自动对车险图像进行处理，识别车险图 像所属的分类。利用本说明书实施例方案，可以大幅提高车险图像分类的准确性和车险图像标注效率，减少人工识别处理耗时，车险图像处理的准确性和可靠性更高。
尽管说明书实施例内容中提到车险图像获取方式、使用场景的分类以及其下类别标签的分类设置、基于深度卷积网络的图像分类算法、基于传统图像特征的分类方式、共享卷积层的神经卷积网络模型等之类的数据模型构建、数据定义、获取、交互、计算、判断等描述，但是，本说明书实施例并不局限于必须是符合行业通信标准、标准计算机数据处理和存储规则或本说明书一个或多个实施例所描述的情况。某些行业标准或者使用自定义方式或实施例描述的实施基础上略加修改后的实施方案也可以实现上述实施例相同、等同或相近、或变形后可预料的实施效果。应用这些修改或变形后的数据获取、存储、判断、处理方式等获取的实施例，仍然可以属于本说明书实施例的可选实施方案范围之内。
上述对本说明书特定实施例进行了描述。其它实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的结果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连续顺序才能实现期望的结果。在某些实施方式中，多任务处理和并行处理也是可以的或者可能是有利的
在20世纪90年代，对于一个技术的改进可以很明显地区分是硬件上的改进(例如，对二极管、晶体管、开关等电路结构的改进)还是软件上的改进(对于方法流程的改进)。然而，随着技术的发展，当今的很多方法流程的改进已经可以视为硬件电路结构的直接改进。设计人员几乎都通过将改进的方法流程编程到硬件电路中来得到相应的硬件电路结构。因此，不能说一个方法流程的改进就不能用硬件实体模块来实现。例如，可编程逻辑器件(Programmable Logic Device,PLD)(例如现场可编程门阵列(Field Programmable Gate Array，FPGA)) 就是这样一种集成电路，其逻辑功能由用户对器件编程来确定。由设计人员自行编程来把一个数字系统“集成”在一片PLD上，而不需要请芯片制造厂商来设计和制作专用的集成电路芯片。而且，如今，取代手工地制作集成电路芯片，这种编程也多半改用“逻辑编译器(logic compiler)”软件来实现，它与程序开发撰写时所用的软件编译器相类似，而要编译之前的原始代码也得用特定的编程语言来撰写，此称之为硬件描述语言(Hardware Description Language，HDL)，而HDL也并非仅有一种，而是有许多种，如ABEL(Advanced Boolean Expression Language)、AHDL(Altera Hardware Description Language)、Confluence、CUPL(Cornell University Programming Language)、HDCal、JHDL(Java Hardware Description Language)、Lava、Lola、MyHDL、PALASM、RHDL(Ruby Hardware Description Language)等，目前最普遍使用的是VHDL(Very-High-Speed Integrated Circuit Hardware Description Language)与Verilog。本领域技术人员也应该清楚，只需要将方法流程用上述几种硬件描述语言稍作逻辑编程并编程到集成电路中，就可以很容易得到实现该逻辑方法流程的硬件电路。
控制器可以按任何适当的方式实现，例如，控制器可以采取例如微处理器或处理器以及存储可由该(微)处理器执行的计算机可读程序代码(例如软件或固件)的计算机可读介质、逻辑门、开关、专用集成电路(Application Specific Integrated Circuit，ASIC)、可编程逻辑控制器和嵌入微控制器的形式，控制器的例子包括但不限于以下微控制器：ARC 625D、Atmel AT91SAM、Microchip PIC18F26K20以及Silicone Labs C8051F320，存储器控制器还可以被实现为存储器的控制逻辑的一部分。本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件部件，而对其内包括的用于实现各种功能的装置也可以视为硬件部件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件部 件内的结构。
上述实施例阐明的系统、装置、模块或单元，具体可以由计算机芯片或实体实现，或者由具有某种功能的产品来实现。一种典型的实现设备为计算机。具体的，计算机例如可以为个人计算机、膝上型计算机、车载人机交互设备、蜂窝电话、相机电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任何设备的组合。
虽然本说明书一个或多个实施例提供了如实施例或流程图所述的方法操作步骤，但基于常规或者无创造性的手段可以包括更多或者更少的操作步骤。实施例中列举的步骤顺序仅仅为众多步骤执行顺序中的一种方式，不代表唯一的执行顺序。在实际中的装置或终端产品执行时，可以按照实施例或者附图所示的方法顺序执行或者并行执行(例如并行处理器或者多线程处理的环境，甚至为分布式数据处理环境)。术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、产品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、产品或者设备所固有的要素。在没有更多限制的情况下，并不排除在包括所述要素的过程、方法、产品或者设备中还存在另外的相同或等同要素。
为了描述的方便，描述以上装置时以功能分为各种模块分别描述。当然，在实施本说明书一个或多个时可以把各模块的功能在同一个或多个软件和/或硬件中实现，也可以将实现同一功能的模块由多个子模块或子单元的组合实现等。以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以 外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件部件，而对其内部包括的用于实现各种功能的装置也可以视为硬件部件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件部件内的结构。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任 何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储、石墨烯存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
本领域技术人员应明白，本说明书一个或多个实施例可提供为方法、系统或计算机程序产品。因此，本说明书一个或多个实施例可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本说明书一个或多个实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本说明书一个或多个实施例可以在由计算机执行的计算机可执行指令的一般上下文中描述，例如程序模块。一般地，程序模块包括执行特定任务或实现特定抽象数据类型的例程、程序、对象、组件、数据结构等等。也可以在分布式计算环境中实践本本说明书一个或多个实施例，在这些分布式计算环境中，由通过通信网络而被连接的远程处理设备来执行任务。在分布式计算环境中，程序模块可以位于包括存储设备在内的本地和远程计算机存储介质中。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于系统实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描 述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本说明书的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述不必须针对的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任一个或多个实施例或示例中以合适的方式结合。此外，在不相互矛盾的情况下，本领域的技术人员可以将本说明书中描述的不同实施例或示例以及不同实施例或示例的特征进行结合和组合。
以上所述仅为本说明书一个或多个实施例的实施例而已，并不用于限制本本说明书一个或多个实施例。对于本领域技术人员来说，本说明书一个或多个实施例可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在权利要求范围之内。

Claims (14)

  1. 一种车险图像处理方法，所述方法包括：
    获取车险图像；
    利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
    按照所述分类标签将所述车险图像进行使用场景的分类存储。
  2. 如权利要求1所述的一种车险图像处理方法，所述预置的图像分类算法包括：
    采用共享卷积层的多任务深度卷积神经网络分类模型。
  3. 如权利要求1或2所述的一种车险图像处理方法，所述的分类存储包括：
    采用关系型数据库存储车险图像的分类结果。
  4. 如权利要求3所述的一种车险图像处理方法，还包括：
    使用选取的光学字符识别算法对所述车险图像进行检测，识别所述车险图像中的文本信息；
    将所述文本信息与所述车险图像进行关联存储。
  5. 如权利要求4所述的一种车险图像处理方法，在识别出所述别所述车险图像中的文本信息之后，所述方法还包括：
    检测所述文本信息中的预设类型的关键信息是否完整，以及记录所述关键信息的检测结果。
  6. 一种车险图像处理装置，所述装置包括：
    图像获取模块，用于获取车险图像；
    图像处理模块，用于利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
    分类存储模块，用于按照所述分类标签将所述车险图像进行使用场景的分类存储。
  7. 如权利要求6所述的一种车险图像处理装置，所述图像处理模块包括：
    多任务分类模型模块，用于采用共享卷积层的多任务深度卷积神经网络分类模型作为所述图像分类算法。
  8. 如权利要求6或7所述的一种车险图像处理装置，所述分类存储模块采用关系型数据库存储车险图像的分类结果。
  9. 如权利要求8所述的一种车险图像处理装置，所述装置还包括：
    文本识别模块，用于使用选取的光学字符识别算法对所述车险图像进行检测，识别所述车险图像中的文本信息；
    文本信息存储模块，用于将所述文本信息与所述车险图像进行关联存储。
  10. 如权利要求9所述的一种车险图像处理装置，所述装置还包括：
    文本信息检测模块，用于检测所述文本信息中的预设类型的关键信息是否完整，以及记录所述关键信息的检测结果。
  11. 一种车险图像处理装置，包括处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
    获取车险图像；
    利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
    按照所述分类标签将所述车险图像进行使用场景的分类存储。
  12. 如权利要求11所述的一种车险图像处理装置，所述处理器执行所述指令时执行的图像分类算法包括：
    采用共享卷积层的多任务深度卷积神经网络分类模型。
  13. 一种服务器，包括至少一个处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
    获取车险图像；
    利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
    按照所述分类标签将所述车险图像进行使用场景的分类存储。
  14. 一种车险图像系统，包括采集图像存储单元、算法服务器、车险图像数据库，所述算法服务器包括至少一个处理器以及用于存储处理器可执行指令的存储器，所述处理器执行所述指令时实现：
    从所述采集图像存储单元获取车险图像；
    利用预置的图像分类算法对所述车险图像进行处理，确定所述车险图像的至少一个类别标签；
    将所述车险图像案子所述类别标签存储到所述车险图像数据库相应的使用场景的存储区域。
PCT/CN2018/097336 2017-07-31 2018-07-27 车险图像处理方法、装置、服务器及系统 Ceased WO2019024771A1 (zh)

Priority Applications (7)

Application Number Priority Date Filing Date Title
MYPI2019005049A MY202176A (en) 2017-07-31 2018-07-27 Vehicle insurance image processing method, apparatus, server, and system
EP18840973.4A EP3621029A4 (en) 2017-07-31 2018-07-27 METHOD, DEVICE, SERVER AND SYSTEM FOR VEHICLE INSURANCE IMAGE PROCESSING
KR1020197027479A KR102373251B1 (ko) 2017-07-31 2018-07-27 차량 보험 이미지 처리 방법, 장치, 서버 및 시스템
JP2019550188A JP7118082B2 (ja) 2017-07-31 2018-07-27 車両保険画像処理方法、装置、サーバ、及びシステム
SG11201908245Q SG11201908245QA (en) 2017-07-31 2018-07-27 Vehicle insurance image processing method, apparatus, server, and system
PH12019502055A PH12019502055A1 (en) 2017-07-31 2019-09-10 Vehicle insurance image processing method, apparatus, server, and system
US16/719,329 US10846556B2 (en) 2017-07-31 2019-12-18 Vehicle insurance image processing method, apparatus, server, and system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710640784.8 2017-07-31
CN201710640784.8A CN107610091A (zh) 2017-07-31 2017-07-31 车险图像处理方法、装置、服务器及系统

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/719,329 Continuation US10846556B2 (en) 2017-07-31 2019-12-18 Vehicle insurance image processing method, apparatus, server, and system

Publications (1)

Publication Number Publication Date
WO2019024771A1 true WO2019024771A1 (zh) 2019-02-07

Family

ID=61064061

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/097336 Ceased WO2019024771A1 (zh) 2017-07-31 2018-07-27 车险图像处理方法、装置、服务器及系统

Country Status (10)

Country Link
US (1) US10846556B2 (zh)
EP (1) EP3621029A4 (zh)
JP (1) JP7118082B2 (zh)
KR (1) KR102373251B1 (zh)
CN (1) CN107610091A (zh)
MY (1) MY202176A (zh)
PH (1) PH12019502055A1 (zh)
SG (1) SG11201908245QA (zh)
TW (1) TWI709919B (zh)
WO (1) WO2019024771A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109948804A (zh) * 2019-03-15 2019-06-28 北京清瞳时代科技有限公司 跨平台拖拽式深度学习建模与训练方法及装置
CN111160342A (zh) * 2019-12-30 2020-05-15 上海眼控科技股份有限公司 车辆反光标识的验证方法、装置、计算机设备和存储介质
CN111353520A (zh) * 2019-07-03 2020-06-30 孙彤 一种朱墨时序鉴定的方法
CN111507850A (zh) * 2020-03-25 2020-08-07 上海商汤智能科技有限公司 核保方法及相关装置、设备
JP2021103565A (ja) * 2020-06-30 2021-07-15 ベイジン バイドゥ ネットコム サイエンス テクノロジー カンパニー リミテッドＢｅｉｊｉｎｇ Ｂａｉｄｕ Ｎｅｔｃｏｍ Ｓｃｉｅｎｃｅ Ｔｅｃｈｎｏｌｏｇｙ Ｃｏ．， Ｌｔｄ． 車両再識別の方法、装置、機器及び記憶媒体

Families Citing this family (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220036463A1 (en) 2012-08-16 2022-02-03 Allstate Insurance Company User devices in claims damage estimation
US10580075B1 (en) * 2012-08-16 2020-03-03 Allstate Insurance Company Application facilitated claims damage estimation
US11532048B2 (en) 2012-08-16 2022-12-20 Allstate Insurance Company User interactions in mobile damage assessment and claims processing
US10430885B1 (en) 2012-08-16 2019-10-01 Allstate Insurance Company Processing insured items holistically with mobile damage assessment and claims processing
US10783585B1 (en) 2012-08-16 2020-09-22 Allstate Insurance Company Agent-facilitated claims damage estimation
US12175540B2 (en) 2012-08-16 2024-12-24 Allstate Insurance Company Processing insured items holistically with mobile damage assessment and claims processing
US8712893B1 (en) 2012-08-16 2014-04-29 Allstate Insurance Company Enhanced claims damage estimation using aggregate display
US8510196B1 (en) 2012-08-16 2013-08-13 Allstate Insurance Company Feedback loop in mobile damage assessment and claims processing
US11455691B2 (en) 2012-08-16 2022-09-27 Allstate Insurance Company Processing insured items holistically with mobile damage assessment and claims processing
CN107610091A (zh) * 2017-07-31 2018-01-19 阿里巴巴集团控股有限公司 车险图像处理方法、装置、服务器及系统
CN109033940B (zh) * 2018-06-04 2019-07-05 上海依图网络科技有限公司 一种图像识别方法、装置、计算设备及存储介质
CN108846766A (zh) * 2018-06-25 2018-11-20 江苏汉德天坤数字技术有限公司 基于深度学习的车险自助理赔成功率预测方法
CN108885699B (zh) * 2018-07-11 2020-06-26 深圳前海达闼云端智能科技有限公司 字符识别方法、装置、存储介质及电子设备
CN110569692B (zh) * 2018-08-16 2023-05-12 创新先进技术有限公司 多车识别方法、装置及设备
CN110569864A (zh) * 2018-09-04 2019-12-13 阿里巴巴集团控股有限公司 基于gan网络的车损图像生成方法和装置
CN109272504B (zh) * 2018-10-17 2022-04-29 广汽丰田汽车有限公司 车辆凹凸缺陷的检测、追溯方法、装置及系统
CN109543773A (zh) * 2018-12-12 2019-03-29 泰康保险集团股份有限公司 图像处理方法、装置、介质及电子设备
CN110033011A (zh) * 2018-12-14 2019-07-19 阿里巴巴集团控股有限公司 车祸事故处理方法和装置、电子设备
CN109903172A (zh) * 2019-01-31 2019-06-18 阿里巴巴集团控股有限公司 理赔信息提取方法和装置、电子设备
CN110674332B (zh) * 2019-08-01 2022-11-15 南昌市微轲联信息技术有限公司 基于ocr与文本挖掘的机动车数字电子档案分类方法
CN110659721B (zh) * 2019-08-02 2022-07-22 杭州未名信科科技有限公司 一种目标检测网络的构建方法和系统
CN110705405B (zh) * 2019-09-20 2021-04-20 创新先进技术有限公司 目标标注的方法及装置
CN111160289A (zh) * 2019-12-31 2020-05-15 欧普照明股份有限公司 针对目标用户的意外事故的检测方法、装置及电子设备
CN111193868B (zh) * 2020-01-09 2021-03-16 中保车服科技服务股份有限公司 车险查勘的拍照方法、装置、计算机设备和可读存储介质
US10818042B1 (en) * 2020-01-14 2020-10-27 Capital One Services, Llc Vehicle information photo overlay
US10832400B1 (en) * 2020-01-14 2020-11-10 Capital One Services, Llc Vehicle listing image detection and alert system
CN111401438B (zh) * 2020-03-13 2023-08-25 德联易控科技(北京)有限公司 图像分拣方法、装置及系统
CN111400529B (zh) * 2020-04-14 2024-03-08 蚂蚁财富(上海)金融信息服务有限公司 数据处理方法以及装置
CN113095211B (zh) * 2021-04-08 2023-08-25 第六镜科技(成都)有限公司 一种图像处理方法、系统及电子设备
CN113918718B (zh) * 2021-10-28 2024-07-02 平安银行股份有限公司 基于人工智能的车险用户分类方法、装置、设备及介质
CN114140025B (zh) * 2021-12-13 2025-03-07 之江实验室 面向多模态数据的车险欺诈行为预测系统、方法和装置
CN114187476A (zh) * 2021-12-14 2022-03-15 中国平安财产保险股份有限公司 基于图像分析的车险信息核对方法、装置、设备及介质
CN114550156B (zh) * 2022-02-18 2024-09-10 支付宝(杭州)信息技术有限公司 图像处理方法及装置
CN115810134B (zh) * 2023-02-14 2023-07-21 之江实验室 一种车险反欺诈的图像采集质检方法、系统和装置
CN117094781B (zh) * 2023-08-25 2024-02-09 国任财产保险股份有限公司 一种智能车险理赔处理方法及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160140424A1 (en) * 2014-11-13 2016-05-19 Nec Laboratories America, Inc. Object-centric Fine-grained Image Classification
CN105678622A (zh) * 2016-01-07 2016-06-15 平安科技（深圳）有限公司 车险理赔照片的分析方法及系统
CN106203330A (zh) * 2016-07-08 2016-12-07 西安理工大学 一种基于卷积神经网络的车辆分类方法
CN107610091A (zh) * 2017-07-31 2018-01-19 阿里巴巴集团控股有限公司 车险图像处理方法、装置、服务器及系统

Family Cites Families (39)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7912246B1 (en) * 2002-10-28 2011-03-22 Videomining Corporation Method and system for determining the age category of people based on facial images
JP2007520934A (ja) * 2003-12-24 2007-07-26 ウオーカー ディジタル、エルエルシー 画像を自動的に捕捉し、管理する方法および装置
KR100631684B1 (ko) * 2004-04-09 2006-10-09 엘지전자 주식회사 보험처리 판단 기능을 갖는 이동 통신 시스템 및 그 방법
US7809587B2 (en) * 2004-05-07 2010-10-05 International Business Machines Corporation Rapid business support of insured property using image analysis
PL1854297T3 (pl) * 2005-02-23 2018-01-31 Gatekeeper Inc Urządzenie do kontroli punktu wjazdu, system i sposób
EP1886484B1 (en) * 2005-05-30 2020-01-01 FUJIFILM Corporation Album creating apparatus, album creating method and program
US7921099B2 (en) * 2006-05-10 2011-04-05 Inquira, Inc. Guided navigation system
JP4931115B2 (ja) 2006-06-29 2012-05-16 キヤノンマーケティングジャパン株式会社 情報処理装置、業務システム、制御方法、プログラム
US8856639B1 (en) * 2007-07-24 2014-10-07 United Services Automobile Association (Usaa) Systems and methods for online document sign-up
US10102583B2 (en) * 2008-01-18 2018-10-16 Mitek Systems, Inc. System and methods for obtaining insurance offers using mobile image capture
US9298979B2 (en) * 2008-01-18 2016-03-29 Mitek Systems, Inc. Systems and methods for mobile image capture and content processing of driver's licenses
US20090234678A1 (en) * 2008-03-11 2009-09-17 Arenas Claims Consulting, Inc. Computer systems and methods for assisting accident victims with insurance claims
US8345981B2 (en) * 2009-02-10 2013-01-01 Kofax, Inc. Systems, methods, and computer program products for determining document validity
JP4743312B2 (ja) * 2009-07-29 2011-08-10 株式会社デンソー 画像識別装置
WO2013093932A2 (en) * 2011-09-29 2013-06-27 Tata Consultancy Services Limited Damage assessment of an object
US8775219B2 (en) * 2012-07-13 2014-07-08 Northrop Grumman Systems Corporation Spectral image classification of rooftop condition for use in property insurance
US8712893B1 (en) * 2012-08-16 2014-04-29 Allstate Insurance Company Enhanced claims damage estimation using aggregate display
US9158744B2 (en) * 2013-01-04 2015-10-13 Cognizant Technology Solutions India Pvt. Ltd. System and method for automatically extracting multi-format data from documents and converting into XML
US9082015B2 (en) * 2013-03-15 2015-07-14 State Farm Mutual Automobile Insurance Company Automatic building assessment
US10319035B2 (en) 2013-10-11 2019-06-11 Ccc Information Services Image capturing and automatic labeling system
CN105960647B (zh) * 2014-05-29 2020-06-09 北京旷视科技有限公司 紧凑人脸表示
US9697469B2 (en) * 2014-08-13 2017-07-04 Andrew McMahon Method and system for generating and aggregating models based on disparate data from insurance, financial services, and public industries
KR101599628B1 (ko) * 2014-08-29 2016-03-04 정유철 교통사고영상 제보시스템
US9659384B2 (en) * 2014-10-03 2017-05-23 EyeEm Mobile GmbH. Systems, methods, and computer program products for searching and sorting images by aesthetic quality
US10387773B2 (en) * 2014-10-27 2019-08-20 Ebay Inc. Hierarchical deep convolutional neural network for image classification
US9524450B2 (en) 2015-03-04 2016-12-20 Accenture Global Services Limited Digital image processing using convolutional neural networks
US9734436B2 (en) * 2015-06-05 2017-08-15 At&T Intellectual Property I, L.P. Hash codes for images
TWI607323B (zh) * 2015-06-05 2017-12-01 永奕科技股份有限公司 車輛資料整合之方法及系統
GB201517462D0 (en) 2015-10-02 2015-11-18 Tractable Ltd Semi-automatic labelling of datasets
JP2017116998A (ja) 2015-12-21 2017-06-29 セゾン自動車火災保険株式会社 情報処理装置、情報処理システム、情報処理方法、情報処理プログラム
JP2017129933A (ja) 2016-01-18 2017-07-27 株式会社ブロードリーフ 画像管理システム、及びこれを用いた車両整備システム
CN105719188B (zh) * 2016-01-22 2017-12-26 平安科技（深圳）有限公司 基于多张图片一致性实现保险理赔反欺诈的方法及服务器
US10394835B1 (en) * 2016-01-22 2019-08-27 C. Joseph Rickrode Rapid access information database (RAID) system and method for generalized data aggregation for a plethora of data types and users
CN106127248A (zh) * 2016-06-24 2016-11-16 平安科技（深圳）有限公司 基于深度学习的车牌分类方法及系统
GB2554361B8 (en) * 2016-09-21 2022-07-06 Emergent Network Intelligence Ltd Automatic image based object damage assessment
US11361380B2 (en) * 2016-09-21 2022-06-14 Allstate Insurance Company Enhanced image capture and analysis of damaged tangible objects
US10902525B2 (en) * 2016-09-21 2021-01-26 Allstate Insurance Company Enhanced image capture and analysis of damaged tangible objects
CN106599869B (zh) * 2016-12-22 2019-12-03 安徽大学 一种基于多任务卷积神经网络的车辆属性识别方法
CN111914692B (zh) * 2017-04-28 2023-07-14 创新先进技术有限公司 车辆定损图像获取方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160140424A1 (en) * 2014-11-13 2016-05-19 Nec Laboratories America, Inc. Object-centric Fine-grained Image Classification
CN105678622A (zh) * 2016-01-07 2016-06-15 平安科技（深圳）有限公司 车险理赔照片的分析方法及系统
CN106203330A (zh) * 2016-07-08 2016-12-07 西安理工大学 一种基于卷积神经网络的车辆分类方法
CN107610091A (zh) * 2017-07-31 2018-01-19 阿里巴巴集团控股有限公司 车险图像处理方法、装置、服务器及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3621029A4 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109948804A (zh) * 2019-03-15 2019-06-28 北京清瞳时代科技有限公司 跨平台拖拽式深度学习建模与训练方法及装置
CN111353520A (zh) * 2019-07-03 2020-06-30 孙彤 一种朱墨时序鉴定的方法
CN111160342A (zh) * 2019-12-30 2020-05-15 上海眼控科技股份有限公司 车辆反光标识的验证方法、装置、计算机设备和存储介质
CN111507850A (zh) * 2020-03-25 2020-08-07 上海商汤智能科技有限公司 核保方法及相关装置、设备
JP2021103565A (ja) * 2020-06-30 2021-07-15 ベイジン バイドゥ ネットコム サイエンス テクノロジー カンパニー リミテッドＢｅｉｊｉｎｇ Ｂａｉｄｕ Ｎｅｔｃｏｍ Ｓｃｉｅｎｃｅ Ｔｅｃｈｎｏｌｏｇｙ Ｃｏ．， Ｌｔｄ． 車両再識別の方法、装置、機器及び記憶媒体
JP7268076B2 (ja) 2020-06-30 2023-05-02 ベイジン バイドゥ ネットコム サイエンス テクノロジー カンパニー リミテッド 車両再識別の方法、装置、機器及び記憶媒体
US11694436B2 (en) 2020-06-30 2023-07-04 Beijing Baidu Netcom Science Technology Co., Ltd. Vehicle re-identification method, apparatus, device and storage medium

Also Published As

Publication number Publication date
MY202176A (en) 2024-04-13
EP3621029A4 (en) 2021-03-03
PH12019502055A1 (en) 2020-07-13
KR20200014725A (ko) 2020-02-11
CN107610091A (zh) 2018-01-19
SG11201908245QA (en) 2019-10-30
US20200125885A1 (en) 2020-04-23
JP2020525871A (ja) 2020-08-27
TW201911129A (zh) 2019-03-16
US10846556B2 (en) 2020-11-24
EP3621029A1 (en) 2020-03-11
KR102373251B1 (ko) 2022-03-14
JP7118082B2 (ja) 2022-08-15
TWI709919B (zh) 2020-11-11

Similar Documents

Publication Publication Date Title
TWI709919B (zh) 車險圖像處理方法、裝置、伺服器及系統
CN107392218B (zh) 一种基于图像的车辆定损方法、装置及电子设备
EP3844668B1 (en) System and method for training a damage identification model
US20200364802A1 (en) Processing method, processing apparatus, user terminal and server for recognition of vehicle damage
KR102151365B1 (ko) 이미지 기반 차량 손실 평가 방법, 장치 및 시스템, 및 전자 디바이스
CN111368600B (zh) 遥感图像目标检测识别方法、装置、可读存储介质及设备
CN110689037A (zh) 用于使用深度网络的自动对象注释的方法和系统
CN107403424A (zh) 一种基于图像的车辆定损方法、装置及电子设备
WO2019214321A1 (zh) 车辆损伤识别的处理方法、处理设备、客户端及服务器
CN113486856A (zh) 一种基于语义分割和卷积神经网络的驾驶员不规范行为检测方法
WO2025152333A1 (zh) 一种缺陷检测方法、装置、设备及存储介质
CN116109597A (zh) 图像篡改区域检测方法、装置、电子设备及存储介质
CN111523439A (zh) 一种基于深度学习的目标检测的方法、系统、设备及介质
US20200050899A1 (en) Automatically filtering out objects based on user preferences
CN114299424B (zh) 目标归类方法、装置、电子设备及存储介质
US12165387B2 (en) Fiber placement tow end detection using machine learning
CN112967187B (zh) 用于目标检测的方法和装置
HK1249264A1 (zh) 车险图像处理方法、装置、服务器及系统
CN118887706B (zh) 一种图像标注方法、装置、电子设备及存储介质
CN115705681A (zh) 城管事件的识别方法、装置、电子设备及存储介质
HK1247374B (zh) 一种基於图像的车辆定损方法、装置及电子设备
CN120148104A (zh) 一种面向职业教育实训车间的学生违规行为检测方法
HK1247423B (zh) 一种基於图像的车辆定损方法、装置及电子设备
CN118506143A (zh) 一种多模态目标检测方法、系统、存储介质及电子设备
HK1246477B (zh) 一种基於图像的车辆定损方法、装置、电子设备及系统

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: 18840973

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019550188

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 20197027479

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2018840973

Country of ref document: EP

Effective date: 20191206

NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Ref document number: 2018840973

Country of ref document: EP