WO2021151313A1 - 证件鉴伪方法、装置、电子设备及存储介质 - Google Patents

证件鉴伪方法、装置、电子设备及存储介质 Download PDF

Info

Publication number
WO2021151313A1
WO2021151313A1 PCT/CN2020/122052 CN2020122052W WO2021151313A1 WO 2021151313 A1 WO2021151313 A1 WO 2021151313A1 CN 2020122052 W CN2020122052 W CN 2020122052W WO 2021151313 A1 WO2021151313 A1 WO 2021151313A1
Authority
WO
WIPO (PCT)
Prior art keywords
authentication
result
certificate
model
probability value
Prior art date
Application number
PCT/CN2020/122052
Other languages
English (en)
French (fr)
Inventor
孟桂国
罗天文
张国辉
宋晨
Original Assignee
平安科技(深圳)有限公司
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 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2021151313A1 publication Critical patent/WO2021151313A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L9/00Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
    • H04L9/32Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials
    • H04L9/3263Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials involving certificates, e.g. public key certificate [PKC] or attribute certificate [AC]; Public key infrastructure [PKI] arrangements
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/80Recognising image objects characterised by unique random patterns
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/40Extraction of image or video features
    • G06V10/44Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/95Pattern authentication; Markers therefor; Forgery detection

Definitions

  • This application relates to the field of artificial intelligence technology, and in particular to a certificate authentication method, device, electronic equipment, and computer-readable storage medium.
  • computers can easily and accurately imitate the human visual perception system for authenticating and identifying documents.
  • the existing certificate authentication methods can be carried out through image processing and traditional machine learning methods.
  • a certificate authentication method provided by this application includes:
  • Data fusion is performed on the multiple prediction results to obtain a result probability value, and the authentication result of the authentication point is obtained according to the result probability value.
  • This application also provides a certificate authentication device, which includes:
  • the image acquisition module is used to acquire the multi-image combination or video frame sequence of the authentication points in the certificate to obtain the set of forged images to be authenticated;
  • the model authentication module is configured to use a pre-trained certificate authentication model to authenticate each picture in the set of images to be authenticated to obtain multiple prediction results;
  • the authentication result output module is used for data fusion of the multiple prediction results to obtain the result probability value, and obtain the authentication result of the authentication point according to the result probability value.
  • This application also provides an electronic device, which includes:
  • Memory storing at least one instruction
  • the processor executes the instructions stored in the memory to implement the following steps:
  • Data fusion is performed on the multiple prediction results to obtain a result probability value, and the authentication result of the authentication point is obtained according to the result probability value.
  • the present application also provides a computer-readable storage medium in which at least one instruction is stored, and the at least one instruction is executed by a processor in an electronic device to implement the following steps:
  • Data fusion is performed on the multiple prediction results to obtain a result probability value, and the authentication result of the authentication point is obtained according to the result probability value.
  • FIG. 1 is a schematic flowchart of a certificate authentication method provided by an embodiment of this application
  • FIG. 2 is a schematic flowchart of a picture detection method provided by an embodiment of this application.
  • FIG. 3 is a schematic flowchart of a method for obtaining a probability value of a result provided by an embodiment of the application
  • FIG. 4 is a schematic diagram of a module of a credential authentication device provided by an embodiment of this application.
  • FIG. 5 is a schematic diagram of the internal structure of an electronic device that implements a method for authenticating a certificate according to an embodiment of the application;
  • the execution subject of the certificate authentication method provided in the embodiment of the present application includes, but is not limited to, at least one of the electronic devices that can be configured to execute the method provided in the embodiment of the present application, such as a server and a terminal.
  • the certificate authentication method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform.
  • the server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.
  • the core of this application is to use computer vision to perform comprehensive authentication judgments on the authentication points in the document under different visual ranges or different illumination conditions, and obtain the authentication result of the document.
  • the authentication point described in the embodiment of the application refers to an anti-counterfeiting mark on the certificate, such as an anti-counterfeiting mark on an ID card.
  • a document may contain multiple authentication points, and the types of authentication points include visual perception authentication points, tactile perception authentication points, and other types of perceptual authentication points.
  • the visual perception authentication point refers to the authentication point that recognizes the authenticity through vision. For example, with different viewing angles, the image of the authentication point will have color replacement, and display waves and three-dimensional effects.
  • the tactile perceptual authentication point refers to the authentication point that recognizes the authenticity by touch, such as touching the authentication point with a finger, which will have a raised feeling; the other types of perceptual authentication points refer to other than visual and tactile authentication points.
  • Ways to identify the authenticity of the authenticity point such as the use of instruments, ultraviolet optometry, and the use of a magnifying glass to identify the micrograph at the authenticity point.
  • the embodiments of the present application are suitable for the authentication judgment of the visual perception authentication point.
  • the certificate authentication method includes:
  • the embodiment of the present application obtains multiple pictures taken from the authentication points from different angles or Video segment, the multi-picture combination or video frame sequence is obtained. Therefore, the multi-picture combination in the embodiment of the present application is a plurality of pictures of the same authentication point at different angles, and the plurality of pictures corresponds to a plurality of shooting angles in a one-to-one manner.
  • the video frame sequence is a video frame sequence obtained by converting a video with the effect of changing the counterfeit detection point into a video frame.
  • the embodiment of the present application before obtaining the video frame sequence of the authentication point in the certificate, also needs to analyze the video key frame of the video. Since the image between adjacent frames of the video does not change much, if each frame is analyzed There will be redundancy, so the image method of extracting the key frame I frame, or the method of skipping frame extraction (such as every 10 frames, 20 frames, etc.) method can be used to analyze the key frames of the video to obtain a video frame sequence.
  • the multi-picture combination or video frame sequence can be obtained from a preset database.
  • the multi-picture combination or video frame sequence of the credential authentication point can also be obtained from Obtained from the preset blockchain node.
  • the credential authentication model in this embodiment of the application may be a deep neural network (Deep Neural Networks, DNN) model used for image recognition, classification and other purposes.
  • the DNN model includes an input layer, a volume Build-up layer, pooling layer, activation layer and output layer.
  • the input layer is used to receive data; the convolutional layer is used to initially extract features from the data; the pooling layer is used to extract main features from the data; the activation layer is used to perform Predictive recognition; the output layer is used to output predictive recognition results.
  • the use of a pre-trained credential authentication model to detect each picture in the to-be-authenticated image set includes:
  • the prediction result is the probability value of the authenticating point in the picture being true.
  • the set of images to be authenticated includes: authentic authentication point images and forged authentication point images.
  • the certificate is used Before extracting the features in the picture by the convolutional layer and the pooling layer of the fake authentication model, it also includes:
  • the convolutional layer and the pooling layer of the document authentication model are improved according to the different characteristics.
  • the counterfeit authentication point is a counterfeit authentication point forged by using methods such as tampering, copying, splicing, and occlusion.
  • the accuracy of the document authentication model in extracting features can be improved.
  • this application further includes training a credential verification model, and the training process of the credential verification model includes :
  • the loss value is greater than or equal to the preset loss threshold, adjust the parameters of the certificate authentication model, and perform authentication again to obtain the training result;
  • the trained certificate authentication model is obtained.
  • the embodiment of the present application uses the following loss function to calculate the difference between the training result and the preset standard result to obtain the difference value:
  • I the training result
  • Y is the standard result
  • represents the error factor, which is a preset constant
  • N is the total amount of data of the sample data.
  • the data fusion of the multiple prediction results to obtain the result probability value includes:
  • the screening of the sorting result set includes:
  • the data that does not belong to the distribution interval is deleted from the sorting result set.
  • P is the result probability value
  • n is the total number of data in the effective result set
  • y i is the i-th data in the effective result set
  • y' is the average value of all data in the effective result set.
  • the embodiment of the present application judges the result probability value according to a preset authentication point confidence threshold, and obtains the authentication result of the certificate corresponding to the original image set to be authenticated.
  • the obtaining the authentication result of the authentication point according to the result probability value includes:
  • the embodiment of the application can authenticate the certificate through a certificate authentication model, which adopts a multi-image combination or a video frame sequence for comprehensive authentication, which improves the accuracy of the authentication result and reduces errors at the same time. Rate.
  • the embodiment of the application obtains the multi-image combination or video frame sequence of the authentication point in the certificate to obtain the set of forged images to be authenticated.
  • the display coverage of the authentication point can be expanded, and the data of the authentication point can be improved.
  • Utilization rate use the pre-trained certificate authentication model to authenticate each picture in the image set to be authenticated, and obtain multiple prediction results, use the certificate authentication model for authentication, and improve the accuracy of authentication and recognition Accuracy; data fusion is performed on the multiple prediction results to obtain the result probability value, and the authentication result of the authentication point is obtained according to the result probability value, and the deviation is reduced by data fusion of the multiple prediction results Errors and extreme interference ensure the accuracy of the results of forgery. Therefore, the certificate authentication method, device and computer-readable storage medium proposed in this application can achieve the purpose of improving the accuracy of certificate authentication.
  • FIG. 4 it is a diagram of the functional modules of the document authentication device of this application.
  • the credential authentication device 100 described in this application can be installed in an electronic device.
  • the certificate authentication device may include an image acquisition module 101, a model authentication module 102, and an authentication result output module 103.
  • the module described in the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can complete fixed functions, and are stored in the memory of the electronic device.
  • each module/unit is as follows:
  • the image acquisition module 101 is used to acquire a combination of multiple images or a sequence of video frames of authentication points in a certificate to obtain a set of images to be authenticated.
  • the embodiment of this application obtains multiple pictures or video segments taken from the authentication points from different angles to obtain The multi-picture combination or sequence of video frames. Therefore, the multi-picture combination in the embodiment of the present application is a plurality of pictures of the same authentication point at different angles, and the plurality of pictures corresponds to a plurality of shooting angles in a one-to-one manner.
  • the video frame sequence is a video frame sequence obtained by converting a video with the effect of changing the counterfeit detection point into a video frame.
  • the embodiment of the present application before obtaining the video frame sequence of the authentication point in the certificate, also needs to analyze the video key frame of the video. Since the image between adjacent frames of the video does not change much, if each frame is analyzed There will be redundancy, so the image method of extracting the key frame I frame, or the method of skipping frame extraction (such as every 10 frames, 20 frames, etc.) method can be used to analyze the key frames of the video to obtain a video frame sequence.
  • the multi-picture combination or video frame sequence can be obtained from a preset database.
  • the multi-picture combination or video frame sequence of the credential authentication point can also be obtained from Obtained from the preset blockchain node.
  • the model verification module 102 is configured to use a pre-trained certificate verification model to verify each picture in the to-be-verified image set to obtain multiple prediction results.
  • the credential authentication model in this embodiment of the application may be a deep neural network (Deep Neural Networks, DNN) model used for image recognition, classification and other purposes.
  • the DNN model includes an input layer, a volume Build-up layer, pooling layer, activation layer and output layer.
  • the input layer is used to receive data; the convolutional layer is used to initially extract features from the data; the pooling layer is used to extract main features from the data; the activation layer is used to perform Predictive recognition; the output layer is used to output predictive recognition results.
  • the model authentication module 102 when using a pre-trained certificate authentication model to detect each picture in the to-be-authenticated image set, the model authentication module 102 specifically performs the following operations:
  • the feature is calculated by using the activation layer of the document authentication model to obtain the prediction result of the picture.
  • the prediction result is the probability value of the authenticating point in the picture being true.
  • the set of images to be authenticated includes: authentic authentication point images and forged authentication point images.
  • the certificate authentication module 102 is also used to:
  • the convolutional layer and the pooling layer of the document authentication model are improved according to the different characteristics.
  • the counterfeit authentication point is a counterfeit authentication point forged by using methods such as tampering, copying, splicing, and occlusion.
  • the accuracy of the document authentication model in extracting features can be improved.
  • this application may further include training a credential verification model, and the credential verification model is used for:
  • the loss value is greater than or equal to the preset loss threshold, adjust the parameters of the certificate authentication model, and perform authentication again to obtain the training result;
  • the trained certificate authentication model is obtained.
  • the embodiment of the present application uses the following loss function to calculate the difference between the training result and the preset standard result to obtain the difference value:
  • I the training result
  • Y is the standard result
  • represents the error factor, which is a preset constant
  • N is the total amount of data of the sample data.
  • the authentication result output module 103 is configured to perform data fusion on the multiple prediction results to obtain a result probability value, and obtain the authentication result of the authentication point according to the result probability value.
  • the authentication result output module 103 specifically performs the following operations:
  • the data in the effective result set is merged using a preset merging algorithm to obtain the result probability value.
  • the screening of the sorting result set includes:
  • the data that does not belong to the distribution interval is deleted from the sorting result set.
  • P is the result probability value
  • n is the total number of data in the effective result set
  • y_i is the i-th data in the effective result set
  • y ⁇ ' is the average value of all data in the effective result set.
  • the embodiment of the present application judges the result probability value according to a preset authentication point confidence threshold, and obtains the authentication result of the certificate corresponding to the original image set to be authenticated.
  • the authentication result output module specifically performs the following operations:
  • the embodiment of the application can authenticate a certificate through a certificate authentication model, which adopts a multi-image combination or a video frame sequence for comprehensive authentication, which improves the accuracy of the authentication result and reduces errors at the same time. Rate.
  • FIG. 5 it is a schematic diagram of the structure of the electronic device implementing the certificate authentication method according to the present application.
  • the electronic device 1 may include a processor 10, a memory 11, and a bus, and may also include a computer program stored in the memory 11 and running on the processor 10, such as a certificate authentication program 12.
  • the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), magnetic memory, magnetic disk, CD etc.
  • the memory 11 may be an internal storage unit of the electronic device 1 in some embodiments, for example, a mobile hard disk of the electronic device 1.
  • the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart media card (SMC), and a secure digital (Secure Digital) equipped on the electronic device 1. , SD) card, flash card (Flash Card), etc.
  • the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device.
  • the memory 11 can be used not only to store application software and various data installed in the electronic device 1, such as the code of the certificate authentication program 12, etc., but also to temporarily store data that has been output or will be output.
  • the processor 10 may be composed of integrated circuits in some embodiments, for example, may be composed of a single packaged integrated circuit, or may be composed of multiple integrated circuits with the same function or different functions, including one or more Combinations of central processing unit (CPU), microprocessor, digital processing chip, graphics processor, and various control chips, etc.
  • the processor 10 is the control unit of the electronic device, which uses various interfaces and lines to connect the various components of the entire electronic device, and runs or executes programs or modules stored in the memory 11 (for example, executing Certificate authentication program, etc.), and call data stored in the memory 11 to execute various functions of the electronic device 1 and process data.
  • the bus may be a peripheral component interconnect standard (PCI) bus or an extended industry standard architecture (EISA) bus, etc.
  • PCI peripheral component interconnect standard
  • EISA extended industry standard architecture
  • the bus can be divided into address bus, data bus, control bus and so on.
  • the bus is configured to implement connection and communication between the memory 11 and at least one processor 10 and the like.
  • FIG. 5 only shows an electronic device with components. Those skilled in the art can understand that the structure shown in FIG. 5 does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure. Components, or a combination of certain components, or different component arrangements.
  • the electronic device 1 may also include a power source (such as a battery) for supplying power to various components.
  • the power source may be logically connected to the at least one processor 10 through a power management device, thereby controlling power
  • the device implements functions such as charge management, discharge management, and power consumption management.
  • the power supply may also include one or more DC or AC power supplies, recharging devices, power failure detection circuits, power converters or inverters, power status indicators and other arbitrary components.
  • the electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.
  • the electronic device 1 may also include a network interface.
  • the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is usually used in the electronic device 1 Establish a communication connection with other electronic devices.
  • the electronic device 1 may also include a user interface.
  • the user interface may be a display (Display) and an input unit (such as a keyboard (Keyboard)).
  • the user interface may also be a standard wired interface or a wireless interface.
  • the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc.
  • the display can also be appropriately called a display screen or a display unit, which is used to display the information processed in the electronic device 1 and to display a visualized user interface.
  • the certificate authentication program 12 stored in the memory 11 in the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can realize:
  • Data fusion is performed on the multiple prediction results to obtain a result probability value, and the authentication result of the authentication point is obtained according to the result probability value.
  • the integrated module/unit of the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium.
  • the computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory)
  • the computer-readable storage medium may be volatile or non-volatile, and when the computer program is executed by a processor, the steps of the aforementioned certificate authentication method are implemented.
  • modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the functional modules in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the above-mentioned integrated unit may be implemented in the form of hardware, or may be implemented in the form of hardware plus software functional modules.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computer Security & Cryptography (AREA)
  • Multimedia (AREA)
  • Mathematical Physics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Collating Specific Patterns (AREA)

Abstract

本申请涉及人工智能技术,揭露了一种证件鉴伪方法,包括:获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。本申请还涉及区块链技术,证件鉴伪点的多图组合或视频帧序列可存储于区块链中。本申请还揭露一种证件鉴伪装置、电子设备及计算机可读存储介质。本申请可以提高证件鉴伪的准确率。

Description

证件鉴伪方法、装置、电子设备及存储介质
本申请要求于2020年7月30日提交中国专利局、申请号为CN202010753977.6、名称为“证件鉴伪方法、装置、电子设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能技术领域,尤其涉及一种证件鉴伪方法、装置、电子设备及计算机可读存储介质。
背景技术
随着计算机的发展和计算机视觉理论的研究,尤其是深度学习和人工智能的出现,计算机可以很方便、很准确地模仿人类的视觉感知系统进行证件的鉴伪识别。
目前现有的证件鉴伪方式可以通过图像处理和传统的机器学习等方法进行,发明人意识到大多是通过图像或单张图像来进行鉴伪,而证件中很多鉴伪点需要不同角度、不同视觉范围或不同光照等条件才能进行综合鉴伪,导致鉴伪结果的准确率较低。
发明内容
本申请提供的一种证件鉴伪方法,包括:
获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
本申请还提供一种证件鉴伪装置,所述装置包括:
图像获取模块,用于获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
模型鉴伪模块,用于利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
鉴伪结果输出模块,用于将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
本申请还提供一种电子设备,所述电子设备包括:
存储器,存储至少一个指令;及
处理器,执行所述存储器中存储的指令以实现如下步骤:
获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
本申请还提供一种计算机可读存储介质,所述计算机可读存储介质中存储有至少一个指令,所述至少一个指令被电子设备中的处理器执行以实现如下步骤:
获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
附图说明
图1为本申请一实施例提供的证件鉴伪方法的流程示意图;
图2为本申请一实施例提供的图片检测方法的流程示意图;
图3为本申请一实施例提供的结果概率值获取方法的流程示意图;
图4为本申请一实施例提供的证件鉴伪装置的模块示意图;
图5为本申请一实施例提供的实现证件鉴伪方法的电子设备的内部结构示意图;
本申请目的的实现、功能特点及优点将结合实施例,参照附图做进一步说明。
具体实施方式
应当理解,此处所描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。
本申请实施例提供的证件鉴伪方法的执行主体包括但不限于服务端、终端等能够被配置为执行本申请实施例提供的该方法的电子设备中的至少一种。换言之,所述证件鉴伪方法可以由安装在终端设备或服务端设备的软件或硬件来执行,所述软件可以是区块链平台。所述服务端包括但不限于:单台服务器、服务器集群、云端服务器或云端服务器集群等。
本申请核心在于利用计算机视觉对证件中的鉴伪点进行不同视觉范围或不同光照等条件下的综合鉴伪判断,得到所述证件的鉴伪结果。
其中,本申请实施例所述鉴伪点是指证件上具有的防伪标记,如身份证件上的防伪标记等。
一个证件中可能包含多个鉴伪点,所述鉴伪点的类型包括视觉感知鉴伪点、触觉感知鉴伪点和其它类型感知鉴伪点。其中,所述视觉感知鉴伪点是指通过视觉识别真伪的鉴伪点,如随着不同观看角度,鉴伪点的图像会有颜色的置换,并显示波浪及立体效果。所述触觉感知鉴伪点是指通过触觉识别真伪的鉴伪点,如通过手指触摸鉴伪点会有凸起的感觉;所述其它类型感知鉴伪点是指通过视觉和触觉以外的其他方式识别真伪的鉴伪点,如借助仪器,紫外线验光、使用放大镜才能辨认鉴伪点处的缩微文字。本申请实施例适用于对视觉感知鉴伪点的鉴伪判断。
参照图1所示,为本申请一实施例提供的证件鉴伪方法的流程示意图。在本实施例中,证件鉴伪方法包括:
S1、获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集。
如上所述,证件中有的鉴伪点在不同观看角度时,图像会有颜色的置换,或显示波浪及立体效果,所以本申请实施例获取鉴伪点从不同角度下拍摄的多张图片或者视频段,得到所述多图组合或视频帧序列。因此,本申请实施例中所述多图组合是同一个鉴伪点在不同角度下的多张图片,多张所述图片与多个拍摄角度一一对应。所述视频帧序列是将拍摄了鉴伪点变化效果的视频转化为视频帧得到的视频帧序列。
可选地,本申请实施例在获取证件中鉴伪点的视频帧序列之前,还需要对视频进行视频关键帧的解析,由于视频相邻帧间图像的变化不大,若每帧都进行解析会有冗余,因此可以采用抽取关键帧I帧的图像方法,或者跳帧抽取(比如每间隔10帧、20帧等)方法等对所述视频进行关键帧的解析,得到视频帧序列。
详细地,所述多图组合或视频帧序列可以从预设的数据库中获取,为进一步保证上述证件信息的私密和安全性,所述证件鉴伪点的多图组合或视频帧序列也可以从预设的区块链节点中获取。
S2、利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果。
较佳地,本申请实施例中所述证件鉴伪模型可以是用于执行图像识别、分类等目的的一种深度神经网络(Deep Neural Networks,DNN)模型,所述DNN模型包括输入层、卷积层、池化层、激活层以及输出层。所述输入层用于接收数据;所述卷积层用于对所述数 据初步提取特征;所述池化层用于对所述数据提取主要特征;所述激活层用于对所述特征进行预测识别;所述输出层用于输出预测识别结果。
详细地,参阅图2所示,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行检测,包括:
S20、利用所述证件鉴伪模型的卷积层和池化层提取所述图片中的特征;
S21、利用所述证件鉴伪模型的激活层对所述特征进行计算,得到所述图片的预测结果。
其中,所述预测结果是所述图片中鉴伪点为真的概率值。
本申请一可选实施例中,所述待鉴伪图像集包括:真鉴伪点图像和伪造鉴伪点图像,为了提高所述证件鉴伪模型的特征提取的准确性,在利用所述证件鉴伪模型卷积层和池化层提取所述图片中的特征之前,还包括:
采集所述真鉴伪点图像,并将所述真鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到真特征数据;
采集所述伪造鉴伪点图像,并将所述伪造鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到伪造特征数据;
分析所述真特征数据和所述伪造特征数据的差别特征;
根据所述差别特征改进所述证件鉴伪模型的卷积层和池化层。
其中,所述伪造鉴伪点是通过利用篡改、翻拍、拼接和遮挡等方式伪造的鉴伪点。
较佳地,通过增强所述证件鉴伪模型的特征表达,可以提高所述证件鉴伪模型在提取特征时的准确性。
可选地,在利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行检测之前,本申请还包括训练证件鉴伪模型,所述证件鉴伪模型的训练过程包括:
生成有效样本数据和所述有效样本数据对应的标准结果;
将所述有效样本数据输入至证件鉴伪模型进行鉴伪,得到训练结果;
利用预设的损失函数对所述训练结果与标准结果进行损失值计算,得到损失值;
当所述损失值大于或等于预设的损失阈值,调整所述证件鉴伪模型的参数,并重新进行鉴伪,得到训练结果;
当所述损失值小于所述损失阈值,得到训练好的所述证件鉴伪模型。
进一步地,本申请实施例利用如下所述损失函数对所述训练结果与预设的标准结果进行差异计算,得到差异值:
Figure PCTCN2020122052-appb-000001
其中,
Figure PCTCN2020122052-appb-000002
为所述训练结果;Y为所述标准结果;α表示误差因子,为预设常数;N是所述样本数据的数据总量。
S3、将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
详细地,参阅图3所示,所述将所述多个预测结果进行数据融合,得到结果概率值,包括:
S30、对所述多个预测结果进行排序,得到排序结果集;
S31、对所述排序结果集进行筛选,得到有效结果集;
S32、利用预设的合并算法将所述有效结果集中的数据进行合并,得到结果概率值。
其中,所述对所述排序结果集进行筛选包括:
确定所述排序结果集中数据的分布区间;
将不属于所述分布区间的数据从所述排序结果集中删除。
本申请实施例中所述预设的合并算法包括:
Figure PCTCN2020122052-appb-000003
其中,P是所述结果概率值,n是所述有效结果集中的数据总数,y i是所述有效结果集第i个数据,y′是所述有效结果集中的所有数据的平均值。
较佳地,本申请实施例根据预设的鉴伪点置信度阈值对所述结果概率值进行判定,得到所述待鉴伪识别的原始图像集对应证件的鉴伪结果。
详细地,所述根据所述结果概率值得到该所述鉴伪点的鉴伪结果,包括:
将所述结果概率值与预设的置信度阈值进行比较;
在所述结果概率值大于等于所述置信度阈值时,得到所述鉴伪点对应的证件为真的鉴伪结果;
在所述结果概率值小于所述置信度阈值时,得到所述鉴伪点对应的证件为假的鉴伪结果。
本申请实施例可以通过证件鉴伪模型对证件进行鉴伪,所述证件鉴伪模型采用多图组合或视频帧序列进行综合鉴伪的方式,提高了鉴伪结果的准确性,同时降低了错误率。
本申请实施例获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集,通过多图组合和视频帧序列可以扩大鉴伪点的呈现覆盖范围,提高鉴伪点数据的利用率;利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果,使用证件鉴伪模型进行鉴伪,提高鉴伪识别的精度和准确率;将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果,通过对多个预测结果进行数据融合,减少偏差错误和极值干扰,保证了鉴伪结果的准确性。因此本申请提出的证件鉴伪方法、装置及计算机可读存储介质,可以实现提高证件鉴伪准确率的目的。
如图4所示,是本申请证件鉴伪装置的功能模块图。
本申请所述证件鉴伪装置100可以安装于电子设备中。根据实现的功能,所述证件鉴伪装置可以包括图像获取模块101、模型鉴伪模块102和鉴伪结果输出模块103。本发所述模块也可以称之为单元,是指一种能够被电子设备处理器所执行,并且能够完成固定功能的一系列计算机程序段,其存储在电子设备的存储器中。
在本实施例中,关于各模块/单元的功能如下:
所述图像获取模块101,用于获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集。
证件中有的鉴伪点在不同观看角度时,图像会有颜色的置换,或显示波浪及立体效果,所以本申请实施例获取鉴伪点从不同角度下拍摄的多张图片或者视频段,得到所述多图组合或视频帧序列。因此,本申请实施例中所述多图组合是同一个鉴伪点在不同角度下的多张图片,多张所述图片与多个拍摄角度一一对应。所述视频帧序列是将拍摄了鉴伪点变化效果的视频转化为视频帧得到的视频帧序列。
可选地,本申请实施例在获取证件中鉴伪点的视频帧序列之前,还需要对视频进行视频关键帧的解析,由于视频相邻帧间图像的变化不大,若每帧都进行解析会有冗余,因此可以采用抽取关键帧I帧的图像方法,或者跳帧抽取(比如每间隔10帧、20帧等)方法等对所述视频进行关键帧的解析,得到视频帧序列。
详细地,所述多图组合或视频帧序列可以从预设的数据库中获取,为进一步保证上述证件信息的私密和安全性,所述证件鉴伪点的多图组合或视频帧序列也可以从预设的区块链节点中获取。
所述模型鉴伪模块102,用于利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果。
较佳地,本申请实施例中所述证件鉴伪模型可以是用于执行图像识别、分类等目的的 一种深度神经网络(Deep Neural Networks,DNN)模型,所述DNN模型包括输入层、卷积层、池化层、激活层以及输出层。所述输入层用于接收数据;所述卷积层用于对所述数据初步提取特征;所述池化层用于对所述数据提取主要特征;所述激活层用于对所述特征进行预测识别;所述输出层用于输出预测识别结果。
详细地,在利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行检测时,所述模型鉴伪模块102具体执行下述操作:
利用所述证件鉴伪模型的卷积层和池化层提取所述图片中的特征;
利用所述证件鉴伪模型的激活层对所述特征进行计算,得到所述图片的预测结果。
其中,所述预测结果是所述图片中鉴伪点为真的概率值。
本申请一可选实施例中,所述待鉴伪图像集包括:真鉴伪点图像和伪造鉴伪点图像,为了提高所述证件鉴伪模型的特征提取的准确性,在利用所述证件鉴伪模型卷积层和池化层提取所述图片中的特征之前,所述模型鉴伪模块102还用于:
采集所述真鉴伪点图像,并将所述真鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到真特征数据;
采集所述伪造鉴伪点图像,并将所述伪造鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到伪造特征数据;
分析所述真特征数据和所述伪造特征数据的差别特征;
根据所述差别特征改进所述证件鉴伪模型的卷积层和池化层。
其中,所述伪造鉴伪点是通过利用篡改、翻拍、拼接和遮挡等方式伪造的鉴伪点。
较佳地,通过增强所述证件鉴伪模型的特征表达,可以提高所述证件鉴伪模型在提取特征时的准确性。
可选地,在利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行检测之前,本申请还可以包括训练证件鉴伪模型,所述证件鉴伪模型用于:
生成有效样本数据和所述有效样本数据对应的标准结果;
将所述有效样本数据输入至证件鉴伪模型进行鉴伪,得到训练结果;
利用预设的损失函数对所述训练结果与标准结果进行损失值计算,得到损失值;
当所述损失值大于或等于预设的损失阈值,调整所述证件鉴伪模型的参数,并重新进行鉴伪,得到训练结果;
当所述损失值小于所述损失阈值,得到训练好的所述证件鉴伪模型。
进一步地,本申请实施例利用如下所述损失函数对所述训练结果与预设的标准结果进行差异计算,得到差异值:
Figure PCTCN2020122052-appb-000004
其中,
Figure PCTCN2020122052-appb-000005
为所述训练结果;Y为所述标准结果;α表示误差因子,为预设常数;N是所述样本数据的数据总量。
所述鉴伪结果输出模块103,用于将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
详细地,在将所述多个预测结果进行数据融合,得到结果概率值,所述鉴伪结果输出模块103具体执行下述操作:
对所述多个预测结果进行排序,得到排序结果集;
对所述排序结果集进行筛选,得到有效结果集;
利用预设的合并算法将所述有效结果集中的数据进行合并,得到结果概率值。
其中,所述对所述排序结果集进行筛选包括:
确定所述排序结果集中数据的分布区间;
将不属于所述分布区间的数据从所述排序结果集中删除。
本申请实施例中所述预设的合并算法包括:
Figure PCTCN2020122052-appb-000006
其中,P是所述结果概率值,n是所述有效结果集中的数据总数,y_i是所述有效结果集第i个数据,y^'是所述有效结果集中的所有数据的平均值。
较佳地,本申请实施例根据预设的鉴伪点置信度阈值对所述结果概率值进行判定,得到所述待鉴伪识别的原始图像集对应证件的鉴伪结果。
详细地,在根据所述结果概率值得到该所述鉴伪点的鉴伪结果时,所述鉴伪结果输出模块具体执行下述操作:
将所述结果概率值与预设的置信度阈值进行比较;
在所述结果概率值大于等于所述置信度阈值时,得到所述鉴伪点对应的证件为真的鉴伪结果;
在所述结果概率值小于所述置信度阈值时,得到所述鉴伪点对应的证件为假的鉴伪结果。
本申请实施例可以通过证件鉴伪模型对证件进行鉴伪,所述证件鉴伪模型采用多图组合或视频帧序列进行综合鉴伪的方式,提高了鉴伪结果的准确性,同时降低了错误率。
如图5所示,是本申请实现证件鉴伪方法的电子设备的结构示意图。
所述电子设备1可以包括处理器10、存储器11和总线,还可以包括存储在所述存储器11中并可在所述处理器10上运行的计算机程序,如证件鉴伪程序12。
其中,所述存储器11至少包括一种类型的可读存储介质,所述可读存储介质包括闪存、移动硬盘、多媒体卡、卡型存储器(例如:SD或DX存储器等)、磁性存储器、磁盘、光盘等。所述存储器11在一些实施例中可以是电子设备1的内部存储单元,例如该电子设备1的移动硬盘。所述存储器11在另一些实施例中也可以是电子设备1的外部存储设备,例如电子设备1上配备的插接式移动硬盘、智能存储卡(Smart Media Card,SMC)、安全数字(Secure Digital,SD)卡、闪存卡(Flash Card)等。进一步地,所述存储器11还可以既包括电子设备1的内部存储单元也包括外部存储设备。所述存储器11不仅可以用于存储安装于电子设备1的应用软件及各类数据,例如证件鉴伪程序12的代码等,还可以用于暂时地存储已经输出或者将要输出的数据。
所述处理器10在一些实施例中可以由集成电路组成,例如可以由单个封装的集成电路所组成,也可以是由多个相同功能或不同功能封装的集成电路所组成,包括一个或者多个中央处理器(Central Processing unit,CPU)、微处理器、数字处理芯片、图形处理器及各种控制芯片的组合等。所述处理器10是所述电子设备的控制核心(Control Unit),利用各种接口和线路连接整个电子设备的各个部件,通过运行或执行存储在所述存储器11内的程序或者模块(例如执行证件鉴伪程序等),以及调用存储在所述存储器11内的数据,以执行电子设备1的各种功能和处理数据。
所述总线可以是外设部件互连标准(peripheral component interconnect,简称PCI)总线或扩展工业标准结构(extended industry standard architecture,简称EISA)总线等。该总线可以分为地址总线、数据总线、控制总线等。所述总线被设置为实现所述存储器11以及至少一个处理器10等之间的连接通信。
图5仅示出了具有部件的电子设备,本领域技术人员可以理解的是,图5示出的结构并不构成对所述电子设备1的限定,可以包括比图示更少或者更多的部件,或者组合某些部件,或者不同的部件布置。
例如,尽管未示出,所述电子设备1还可以包括给各个部件供电的电源(比如电池),优选地,电源可以通过电源管理装置与所述至少一个处理器10逻辑相连,从而通过电源管理装置实现充电管理、放电管理、以及功耗管理等功能。电源还可以包括一个或一个以 上的直流或交流电源、再充电装置、电源故障检测电路、电源转换器或者逆变器、电源状态指示器等任意组件。所述电子设备1还可以包括多种传感器、蓝牙模块、Wi-Fi模块等,在此不再赘述。
进一步地,所述电子设备1还可以包括网络接口,可选地,所述网络接口可以包括有线接口和/或无线接口(如WI-FI接口、蓝牙接口等),通常用于在该电子设备1与其他电子设备之间建立通信连接。
可选地,该电子设备1还可以包括用户接口,用户接口可以是显示器(Display)、输入单元(比如键盘(Keyboard)),可选地,用户接口还可以是标准的有线接口、无线接口。可选地,在一些实施例中,显示器可以是LED显示器、液晶显示器、触控式液晶显示器以及OLED(Organic Light-Emitting Diode,有机发光二极管)触摸器等。其中,显示器也可以适当的称为显示屏或显示单元,用于显示在电子设备1中处理的信息以及用于显示可视化的用户界面。
应该了解,所述实施例仅为说明之用,在专利申请范围上并不受此结构的限制。
所述电子设备1中的所述存储器11存储的证件鉴伪程序12是多个指令的组合,在所述处理器10中运行时,可以实现:
获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
进一步地,所述电子设备1集成的模块/单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中。所述计算机可读介质可以包括:能够携带所述计算机程序代码的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(ROM,Read-Only Memory),所述计算机可读存储介质可以是易失性,也可以是非易失性,所述计算机程序被处理器执行时实现上述证件鉴伪方法的步骤。
在本申请所提供的几个实施例中,应该理解到,所揭露的设备,装置和方法,可以通过其它的方式实现。例如,以上所描述的装置实施例仅仅是示意性的,例如,所述模块的划分,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。
所述作为分离部件说明的模块可以是或者也可以不是物理上分开的,作为模块显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
另外,在本申请各个实施例中的各功能模块可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现,也可以采用硬件加软件功能模块的形式实现。
对于本领域技术人员而言,显然本申请不限于上述示范性实施例的细节,而且在不背离本申请的精神或基本特征的情况下,能够以其他的具体形式实现本申请。
因此,无论从哪一点来看,均应将实施例看作是示范性的,而且是非限制性的,本申请的范围由所附权利要求而不是上述说明限定,因此旨在将落在权利要求的等同要件的含义和范围内的所有变化涵括在本申请内。不应将权利要求中的任何附关联图表记视为限制所涉及的权利要求。
此外,显然“包括”一词不排除其他单元或步骤,单数不排除复数。系统权利要求中陈述的多个单元或装置也可以由一个单元或装置通过软件或者硬件来实现。第二等词语用来表示名称,而并不表示任何特定的顺序。
最后应说明的是,以上实施例仅用以说明本申请的技术方案而非限制,尽管参照较佳 实施例对本申请进行了详细说明,本领域的普通技术人员应当理解,可以对本申请的技术方案进行修改或等同替换,而不脱离本申请技术方案的精神和范围。

Claims (20)

  1. 一种证件鉴伪方法,其中,所述方法包括:
    获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
    利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
    将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
  2. 如权利要求1所述的证件鉴伪方法,其中,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,包括:
    利用所述证件鉴伪模型的卷积层和池化层提取所述图片中的特征;
    利用所述证件鉴伪模型的激活层对所述特征进行计算,得到所述图片的预测结果。
  3. 如权利要求2所述的证件鉴伪方法,其中,所述待鉴伪图像集包括:真鉴伪点图像和伪造鉴伪点图像,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果之前,该方法还包括:
    采集所述真鉴伪点图像,并将所述真鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到真特征数据;
    采集所述伪造鉴伪点图像,并将所述伪造鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到伪造特征数据;
    分析所述真特征数据和所述伪造特征数据的差别特征;
    根据所述差别特征改进所述证件鉴伪模型的卷积层和池化层。
  4. 如权利要求1所述的证件鉴伪方法,其中,所述将所述多个预测结果进行数据融合,得到结果概率值,包括:
    对所述多个预测结果进行排序,得到排序结果集;
    对所述排序结果集进行筛选,得到有效结果集;
    利用预设的合并算法将所述有效结果集中的数据进行合并,得到结果概率值。
  5. 如权利要求4所述的证件鉴伪方法,其中,所述对所述排序结果集进行筛选包括:
    确定所述排序结果集中数据的分布区间;
    将不属于所述分布区间的数据从所述排序结果集中删除。
  6. 如权利要求1所述的证件鉴伪方法,其中,所述根据所述结果概率值得到该所述鉴伪点的鉴伪结果,包括:
    将所述结果概率值与预设的置信度阈值进行比较;
    在所述结果概率值大于等于所述置信度阈值时,得到所述鉴伪点对应的证件为真的鉴伪结果;
    在所述结果概率值小于所述置信度阈值时,得到所述鉴伪点对应的证件为假的鉴伪结果。
  7. 如权利要求1至6中任意一项所述的证件鉴伪方法,其中,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果之前,该方法还包括:
    生成有效样本数据和所述有效样本数据对应的标准结果;
    将所述有效样本数据输入至证件鉴伪模型进行鉴伪,得到训练结果;
    利用预设的损失函数对所述训练结果与标准结果进行损失值计算,得到损失值;
    当所述损失值大于或等于预设的损失阈值,调整所述证件鉴伪模型的参数,并重新进行鉴伪,得到训练结果;
    当所述损失值小于所述损失阈值,得到训练好的所述证件鉴伪模型。
  8. 一种证件鉴伪装置,其中,所述装置包括:
    图像获取模块,用于获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
    模型鉴伪模块,用于利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
    鉴伪结果输出模块,用于将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
  9. 一种电子设备,其中,所述电子设备包括:
    存储器,存储至少一个指令;及
    处理器,执行所述存储器中存储的指令以执行如下步骤:
    获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
    利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
    将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
  10. 如权利要求9所述的电子设备,其中,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,包括:
    利用所述证件鉴伪模型的卷积层和池化层提取所述图片中的特征;
    利用所述证件鉴伪模型的激活层对所述特征进行计算,得到所述图片的预测结果。
  11. 如权利要求10所述的电子设备,其中,所述待鉴伪图像集包括:真鉴伪点图像和伪造鉴伪点图像,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果之前,所述处理器执行所述指令时还实现如下步骤:
    采集所述真鉴伪点图像,并将所述真鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到真特征数据;
    采集所述伪造鉴伪点图像,并将所述伪造鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到伪造特征数据;
    分析所述真特征数据和所述伪造特征数据的差别特征;
    根据所述差别特征改进所述证件鉴伪模型的卷积层和池化层。
  12. 如权利要求9所述的电子设备,其中,所述将所述多个预测结果进行数据融合,得到结果概率值,包括:
    对所述多个预测结果进行排序,得到排序结果集;
    对所述排序结果集进行筛选,得到有效结果集;
    利用预设的合并算法将所述有效结果集中的数据进行合并,得到结果概率值。
  13. 如权利要求12所述的电子设备,其中,所述对所述排序结果集进行筛选包括:
    确定所述排序结果集中数据的分布区间;
    将不属于所述分布区间的数据从所述排序结果集中删除。
  14. 如权利要求9所述的电子设备,其中,所述根据所述结果概率值得到该所述鉴伪点的鉴伪结果,包括:
    将所述结果概率值与预设的置信度阈值进行比较;
    在所述结果概率值大于等于所述置信度阈值时,得到所述鉴伪点对应的证件为真的鉴伪结果;
    在所述结果概率值小于所述置信度阈值时,得到所述鉴伪点对应的证件为假的鉴伪结果。
  15. 如权利要求9至14中任意一项所述的电子设备,其中,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果之前,所述处理器执行所述指令时还实现如下步骤:
    生成有效样本数据和所述有效样本数据对应的标准结果;
    将所述有效样本数据输入至证件鉴伪模型进行鉴伪,得到训练结果;
    利用预设的损失函数对所述训练结果与标准结果进行损失值计算,得到损失值;
    当所述损失值大于或等于预设的损失阈值,调整所述证件鉴伪模型的参数,并重新进行鉴伪,得到训练结果;
    当所述损失值小于所述损失阈值,得到训练好的所述证件鉴伪模型。
  16. 一种计算机可读存储介质,存储有计算机程序,其中,所述计算机程序被处理器执行时实现如下步骤:
    获取证件中鉴伪点的多图组合或视频帧序列,得到待鉴伪图像集;
    利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果;
    将所述多个预测结果进行数据融合,得到结果概率值,并根据所述结果概率值得到该所述鉴伪点的鉴伪结果。
  17. 如权利要求16所述的计算机可读存储介质,其中,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,包括:
    利用所述证件鉴伪模型的卷积层和池化层提取所述图片中的特征;
    利用所述证件鉴伪模型的激活层对所述特征进行计算,得到所述图片的预测结果。
  18. 如权利要求17所述的计算机可读存储介质,其中,所述待鉴伪图像集包括:真鉴伪点图像和伪造鉴伪点图像,所述利用预先训练好的证件鉴伪模型对所述待鉴伪图像集中的每张图片进行鉴伪,得到多个预测结果之前,所述计算机程序被处理器执行时还实现如下步骤:
    采集所述真鉴伪点图像,并将所述真鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到真特征数据;
    采集所述伪造鉴伪点图像,并将所述伪造鉴伪点图像输入至所述证件鉴伪模型中进行特征提取,得到伪造特征数据;
    分析所述真特征数据和所述伪造特征数据的差别特征;
    根据所述差别特征改进所述证件鉴伪模型的卷积层和池化层。
  19. 如权利要求16所述的计算机可读存储介质,其中,所述将所述多个预测结果进行数据融合,得到结果概率值,包括:
    对所述多个预测结果进行排序,得到排序结果集;
    对所述排序结果集进行筛选,得到有效结果集;
    利用预设的合并算法将所述有效结果集中的数据进行合并,得到结果概率值。
  20. 如权利要求19所述的计算机可读存储介质,其中,所述对所述排序结果集进行筛选包括:
    确定所述排序结果集中数据的分布区间;
    将不属于所述分布区间的数据从所述排序结果集中删除。
PCT/CN2020/122052 2020-07-30 2020-10-20 证件鉴伪方法、装置、电子设备及存储介质 WO2021151313A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010753977.6A CN111898538B (zh) 2020-07-30 2020-07-30 证件鉴伪方法、装置、电子设备及存储介质
CN202010753977.6 2020-07-30

Publications (1)

Publication Number Publication Date
WO2021151313A1 true WO2021151313A1 (zh) 2021-08-05

Family

ID=73184083

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/122052 WO2021151313A1 (zh) 2020-07-30 2020-10-20 证件鉴伪方法、装置、电子设备及存储介质

Country Status (2)

Country Link
CN (1) CN111898538B (zh)
WO (1) WO2021151313A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114495176A (zh) * 2022-03-30 2022-05-13 北京字节跳动网络技术有限公司 组织图像的识别方法、装置、可读介质和电子设备
CN114494765A (zh) * 2021-12-21 2022-05-13 北京瑞莱智慧科技有限公司 真假烟鉴别点的识别方法、装置、电子设备及存储介质
CN114648814A (zh) * 2022-02-25 2022-06-21 北京百度网讯科技有限公司 人脸活体检测方法及模型的训练方法、装置、设备及介质
CN117133039A (zh) * 2023-09-01 2023-11-28 中国科学院自动化研究所 图像鉴伪模型训练方法、图像鉴伪方法、装置及电子设备

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112528998B (zh) * 2021-02-18 2021-06-01 成都新希望金融信息有限公司 证件图像处理方法、装置、电子设备及可读存储介质
CN113240043B (zh) * 2021-06-01 2024-04-09 平安科技(深圳)有限公司 基于多图片差异性的鉴伪方法、装置、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6175644B1 (en) * 1998-05-01 2001-01-16 Cognex Corporation Machine vision system for object feature analysis and validation based on multiple object images
CN110046644A (zh) * 2019-02-26 2019-07-23 阿里巴巴集团控股有限公司 一种证件防伪的方法及装置、计算设备和存储介质
CN110188659A (zh) * 2019-05-27 2019-08-30 Oppo广东移动通信有限公司 健康检测方法及相关产品
CN111324874A (zh) * 2020-01-21 2020-06-23 支付宝实验室(新加坡)有限公司 一种证件真伪识别方法及装置
CN111898520A (zh) * 2020-07-28 2020-11-06 腾讯科技(深圳)有限公司 证件真伪识别方法、装置、计算机可读介质及电子设备

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106127103B (zh) * 2016-06-12 2019-06-25 广州广电运通金融电子股份有限公司 一种离线身份认证的方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6175644B1 (en) * 1998-05-01 2001-01-16 Cognex Corporation Machine vision system for object feature analysis and validation based on multiple object images
CN110046644A (zh) * 2019-02-26 2019-07-23 阿里巴巴集团控股有限公司 一种证件防伪的方法及装置、计算设备和存储介质
CN110188659A (zh) * 2019-05-27 2019-08-30 Oppo广东移动通信有限公司 健康检测方法及相关产品
CN111324874A (zh) * 2020-01-21 2020-06-23 支付宝实验室(新加坡)有限公司 一种证件真伪识别方法及装置
CN111898520A (zh) * 2020-07-28 2020-11-06 腾讯科技(深圳)有限公司 证件真伪识别方法、装置、计算机可读介质及电子设备

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114494765A (zh) * 2021-12-21 2022-05-13 北京瑞莱智慧科技有限公司 真假烟鉴别点的识别方法、装置、电子设备及存储介质
CN114494765B (zh) * 2021-12-21 2023-08-18 北京瑞莱智慧科技有限公司 真假烟鉴别点的识别方法、装置、电子设备及存储介质
CN114648814A (zh) * 2022-02-25 2022-06-21 北京百度网讯科技有限公司 人脸活体检测方法及模型的训练方法、装置、设备及介质
CN114495176A (zh) * 2022-03-30 2022-05-13 北京字节跳动网络技术有限公司 组织图像的识别方法、装置、可读介质和电子设备
CN114495176B (zh) * 2022-03-30 2022-12-06 北京字节跳动网络技术有限公司 组织图像的识别方法、装置、可读介质和电子设备
CN117133039A (zh) * 2023-09-01 2023-11-28 中国科学院自动化研究所 图像鉴伪模型训练方法、图像鉴伪方法、装置及电子设备
CN117133039B (zh) * 2023-09-01 2024-03-15 中国科学院自动化研究所 图像鉴伪模型训练方法、图像鉴伪方法、装置及电子设备

Also Published As

Publication number Publication date
CN111898538B (zh) 2024-04-12
CN111898538A (zh) 2020-11-06

Similar Documents

Publication Publication Date Title
WO2021151313A1 (zh) 证件鉴伪方法、装置、电子设备及存储介质
CN111886842B (zh) 使用基于阈值的匹配进行远程用户身份验证
WO2021151277A1 (zh) 目标物损伤程度判定方法、装置、电子设备及存储介质
WO2022105179A1 (zh) 生物特征图像识别方法、装置、电子设备及可读存储介质
WO2022247005A1 (zh) 图像中目标物识别方法、装置、电子设备及存储介质
CN109145766A (zh) 模型训练方法、装置、识别方法、电子设备及存储介质
CN107967461B (zh) Svm差分模型训练及人脸验证方法、装置、终端及存储介质
CN113033543B (zh) 曲形文本识别方法、装置、设备及介质
CN113887408B (zh) 活化人脸视频的检测方法、装置、设备及存储介质
WO2021128846A1 (zh) 电子文件的控制方法、装置、计算机设备及存储介质
WO2021189856A1 (zh) 证件校验方法、装置、电子设备及介质
CN114881698A (zh) 广告合规审核方法、装置、电子设备及存储介质
CN115471775A (zh) 基于录屏视频的信息验证方法、装置、设备及存储介质
CN108648189A (zh) 图像模糊检测方法、装置、计算设备及可读存储介质
CN113869219B (zh) 人脸活体检测方法、装置、设备及存储介质
CN112528903B (zh) 人脸图像获取方法、装置、电子设备及介质
CN113705469A (zh) 人脸识别方法、装置、电子设备及计算机可读存储介质
CN114639152A (zh) 基于人脸识别的多模态语音交互方法、装置、设备及介质
CN112037235B (zh) 伤情图片自动化审核方法、装置、电子设备及存储介质
CN113888500A (zh) 基于人脸图像的炫光程度检测方法、装置、设备及介质
CN111275683B (zh) 图像质量评分处理方法、系统、设备及介质
CN117437522A (zh) 一种人脸识别模型训练方法、人脸识别方法及装置
WO2023134080A1 (zh) 相机作弊识别方法、装置、设备及存储介质
CN113255456B (zh) 非主动活体检测方法、装置、电子设备及存储介质
WO2022156088A1 (zh) 指纹签名生成方法、装置、电子设备及计算机存储介质

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: 20917117

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20917117

Country of ref document: EP

Kind code of ref document: A1