Palm print recognition method and device, storage medium and electronic equipment

Abstract

The invention discloses a palmprint recognition method and device, a storage medium and electronic equipment, and belongs to the technical field of artificial intelligence, and the method comprises the steps: obtaining at least one palm image sample of each known user in a plurality of known users; determining a palmprint area image and a first palmprint curve image of the corresponding palmprint according to each palm image sample; according to the palmprint area image and the first palmprint curve image, a feature extraction model is trained to obtain a palmprint recognition model, the feature extraction model comprises a first feature extraction module and a second feature extraction module which have the same network structure and share parameters, the first feature extraction module is used for extracting features of the palmprint area image, and the second feature extraction module is used for extracting features of the palmprint curve image; the second feature extraction module is used for extracting features of the first palmprint curve image; palmprint recognition is performed based on the palmprint recognition model, so that data enhancement training of the palmprint recognition model can be realized, and the palmprint recognition accuracy of the model is improved.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and in particular, relates to a palmprint recognition method, device, storage medium and electronic device. Background Art

[0002] With the widespread application of smart devices, people have increasing requirements for the reliability, security, accuracy, and convenience of biometric technology used for identity recognition. Compared with existing commonly used biometric features, such as voice, face, iris, fingerprint, gait, etc., palm prints have strong self-stability and individual differences, providing a relatively important basis for identity recognition, and thus have been widely valued by many researchers.

[0003] At present, the palmprint recognition method based on deep learning uses a learning model to extract features from palmprint images. However, due to the relatively complex acquisition and annotation of palmprint data, it is difficult to collect large-scale palmprint data samples for model training, resulting in the model being unable to learn and generalize well, which in turn makes it difficult for the trained model to support large-scale palmprint recognition tasks, the extraction accuracy of palmprint features is low, and the recognition effect is poor. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a palmprint recognition method, device, storage medium and electronic device, which can accurately extract palmprint features and improve palmprint recognition effect.

[0005] In a first aspect, the present application provides a palmprint recognition method, comprising:

[0006] obtaining at least one palm image sample of each known user among a plurality of known users;

[0007] Determine a palmprint region image and a first palmprint curve image of a corresponding palmprint according to each of the palm image samples;

[0008] According to the palmprint area image and the first palmprint curve image, a feature extraction model is trained to obtain a palmprint recognition model, wherein the feature extraction model includes a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is used to extract features of the palmprint area image, and the second feature extraction module is used to extract features of the first palmprint curve image;

[0009] Palmprint recognition is performed based on the palmprint recognition model.

[0010] In some embodiments, determining the palm print region image and the first palm print curve image of the corresponding palm print according to each of the palm image samples includes:

[0011] Extracting an image of the area where the corresponding palm print is located from each of the palm image samples to obtain a palm print area image;

[0012] The palmprint region image is input into a trained geometric curve generation model to generate a binary geometric curve image, thereby obtaining a first palmprint curve image corresponding to the palmprint.

[0013] In some embodiments, the training process of the geometric curve generation model includes:

[0014] Generate a binary geometric curve image corresponding to each of the palmprint area images using the constructed geometric curve generation model to obtain a second palmprint curve image;

[0015] Generative adversarial training is performed on the geometric curve generation model according to the palmprint area image and the second palmprint curve image.

[0016] In some embodiments, extracting an image of the area where the corresponding palm print is located from each of the palm image samples includes:

[0017] Performing key point detection on each palm image sample to obtain multiple key points;

[0018] According to the image coordinates of the key points, an image of the area where the corresponding palm print is located is extracted from the palm image sample.

[0019] In some embodiments, the training of the feature extraction model according to the palmprint area image and the first palmprint curve image includes:

[0020] Using the first feature extraction module to perform feature extraction on the palmprint region image to obtain a first extracted feature, and using the second feature extraction module to perform feature extraction on the first palmprint curve image to obtain a second extracted feature;

[0021] The feature extraction model is trained according to a target loss function, the first extracted features and the second extracted features, wherein the target loss function includes a weighted first loss function and a second loss function, wherein the first loss function is used to determine a loss value according to features in the palmprint area image, and the second loss function is used to determine a loss value according to features in the palmprint area image and features in the first palmprint curve image.

[0022] In some embodiments, the training of the feature extraction model according to the target loss function, the first extracted features, and the second extracted features includes:

[0023] Determine a first loss value according to the first extracted feature and the first loss function, and determine a second loss value according to the first extracted feature, the second extracted feature, and the second loss function;

[0024] Determining a comprehensive loss value by weighting according to the first loss value and the second loss value;

[0025] The parameters of the feature extraction model are adjusted according to the comprehensive loss value to train the feature extraction model.

[0026] In some embodiments, performing palmprint recognition based on the palmprint recognition model includes:

[0027] Obtain a target palm image of a user to be identified;

[0028] Determine a target palm print region image and a target palm print curve image of a corresponding palm print according to the target palm image;

[0029] The target palmprint curve image and the target palmprint area image are input into the palmprint recognition model to obtain palmprint feature data of the user to be recognized.

[0030] In a second aspect, the present application provides a palmprint recognition device, comprising:

[0031] an acquisition unit, configured to acquire at least one palm image sample of each known user among a plurality of known users;

[0032] A determination unit, configured to determine a palm print region image and a first palm print curve image of a corresponding palm print according to each of the palm image samples;

[0033] a training unit, configured to train a feature extraction model according to the palmprint region image and the first palmprint curve image to obtain a palmprint recognition model, wherein the feature extraction model comprises a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is configured to extract features of the palmprint region image, and the second feature extraction module is configured to extract features of the first palmprint curve image;

[0034] The recognition unit is used to perform palmprint recognition based on the palmprint recognition model.

[0035] In some embodiments, the determining unit is specifically configured to:

[0036] Extracting an image of the area where the corresponding palm print is located from each of the palm image samples to obtain a palm print area image;

[0037] The palmprint region image is input into a trained geometric curve generation model to generate a binary geometric curve image, thereby obtaining a first palmprint curve image corresponding to the palmprint.

[0038] In some embodiments, the determining unit is specifically configured to:

[0039] Performing key point detection on each palm image sample to obtain multiple key points;

[0040] According to the image coordinates of the key points, an image of the area where the corresponding palm print is located is extracted from the palm image sample.

[0041] In some embodiments, the training unit is further configured to:

[0042] Generate a binary geometric curve image corresponding to each of the palmprint area images using the constructed geometric curve generation model to obtain a second palmprint curve image;

[0043] Generative adversarial training is performed on the geometric curve generation model according to the palmprint area image and the second palmprint curve image.

[0044] In some embodiments, the training unit is specifically used for:

[0045] Using the first feature extraction module to perform feature extraction on the palmprint region image to obtain a first extracted feature, and using the second feature extraction module to perform feature extraction on the first palmprint curve image to obtain a second extracted feature;

[0046] The feature extraction model is trained according to a target loss function, the first extracted features and the second extracted features, wherein the target loss function includes a weighted first loss function and a second loss function, wherein the first loss function is used to determine a loss value according to features in the palmprint area image, and the second loss function is used to determine a loss value according to features in the palmprint area image and features in the first palmprint curve image.

[0047] In some embodiments, the training unit is specifically used for:

[0048] Determine a first loss value according to the first extracted feature and the first loss function, and determine a second loss value according to the first extracted feature, the second extracted feature, and the second loss function;

[0049] Determining a comprehensive loss value by weighting according to the first loss value and the second loss value;

[0050] The parameters of the feature extraction model are adjusted according to the comprehensive loss value to train the feature extraction model.

[0051] In some embodiments, the identification unit is specifically used to:

[0052] Obtain a target palm image of a user to be identified;

[0053] Determine a target palm print region image and a target palm print curve image of a corresponding palm print according to the target palm image;

[0054] The target palmprint curve image and the target palmprint area image are input into the palmprint recognition model to obtain palmprint feature data of the user to be recognized.

[0055] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any of the palmprint recognition methods described above.

[0056] In a fourth aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the above-mentioned palmprint recognition methods when executing the program.

[0057] In a fifth aspect, the present application provides a computer program product, including a computer program, which implements any of the above-mentioned palmprint recognition methods when executed by a processor.

[0058] The palmprint recognition method, device, storage medium, electronic device and computer program product provided by the embodiments of the present application obtain at least one palm image sample of each known user among multiple known users; determine the palmprint area image and the first palmprint curve image of the corresponding palmprint according to each palm image sample; train a feature extraction model according to the palmprint area image and the first palmprint curve image to obtain a palmprint recognition model, the feature extraction model includes a first feature extraction module and a second feature extraction module with the same network structure and shared parameters, the first feature extraction module is used to extract the features of the palmprint area image, and the second feature extraction module is used to extract the features of the first palmprint curve image; palmprint recognition is performed based on the palmprint recognition model, that is, the model is trained using the image of the area where the palmprint is located and the curve image corresponding to the palmprint texture, so that data enhancement training of the palmprint recognition model can be achieved, ensuring that the model can better learn the features of the main line and fine lines of the palmprint, so that the model can be well learned and generalized, thereby improving the recognition accuracy of the model for large-scale palmprint recognition tasks, and the palmprint recognition effect is good. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0060] Figure 1 It is a flowchart of a palmprint recognition method provided in an embodiment of the present application;

[0061] Figure 2 is another schematic diagram of the palmprint recognition method provided by an embodiment of the present application;

[0062] Figure 3 is a schematic diagram showing a first palm print curve image provided in an embodiment of the present application;

[0063] Figure 4 A schematic diagram of a data enhancement training process for a feature extraction model provided in an embodiment of the present application;

[0064] Figure 5 A schematic diagram of the structure of a palmprint recognition device provided in an embodiment of the present application;

[0065] Figure 6 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0066] Figure 7 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0067] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be understood as limiting the present application.

[0068] Embodiments of the present application provide a palmprint recognition method, device, storage medium, electronic device, and computer program product.

[0069] See also Figure 1 , Figure 1 1 is a flow chart of a palmprint recognition method provided by an embodiment of the present application. The palmprint recognition method is applied to an electronic device and can be executed by the electronic device or a functional module or functional entity in the electronic device. The palmprint recognition method includes the following steps 101-104, wherein:

[0070] 101. Obtain at least one palm image sample of each known user among a plurality of known users;

[0071] 102. Determine a palm print region image and a first palm print curve image of a corresponding palm print according to each palm image sample;

[0072] 103. According to the palmprint region image and the first palmprint curve image, a feature extraction model is trained to obtain a palmprint recognition model, wherein the feature extraction model includes a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is used to extract features of the palmprint region image, and the second feature extraction module is used to extract features of the first palmprint curve image;

[0073] 104. Perform palmprint recognition based on the palmprint recognition model.

[0074] The palmprint recognition method provided in the embodiment of the present application obtains a palm image sample of a known user to determine a palmprint area image and a first palmprint curve image, and uses the palmprint area image and the first palmprint curve image to perform model training, that is, uses an image of the area where the palmprint is located and a curve image corresponding to the palmprint texture to train the model, thereby realizing data enhancement training of the palmprint recognition model, ensuring that the model learns the features of the main lines and fine lines of the palmprint well, enabling the model to be well learned and generalized, thereby improving the recognition accuracy of the model for large-scale palmprint recognition tasks, and achieving good palmprint recognition effect.

[0075] The above steps 101 - 104 will be described in detail below.

[0076] 101. Obtain at least one palm image sample of each known user among multiple known users.

[0077] Among them, an image acquisition device such as a camera can be used to capture the palm of each known user. If the captured image is determined to be an image of the palm and passes the palm quality detection algorithm threshold, it will be used as a palm image sample, for example, denoted as G. These images retain the original characteristics of the palm print.

[0078] In some embodiments, since various noises and deformations may be introduced during the image acquisition process, such as environmental noise, equipment noise, geometric deformation, nonlinear deformation, etc., in order to avoid effective features in the image being masked by noise or deformation and affecting the quality of the image sample, preprocessing can be performed before subsequent processing of the image sample. The preprocessing includes but is not limited to denoising and correction.

[0079] For example, denoising the original palm image sample G can effectively remove the noise in the image and obtain a clearer image, which is recorded as G1. Taking Gaussian filter denoising as an example, the formula is: Where x and y are the coordinate components of the pixel in the image, and σ is the standard deviation of the Gaussian function. σ determines the width of the Gaussian function, which in turn affects the smoothness of the filter. The larger the σ value, the flatter the Gaussian curve and the smoother the filtering effect; the smaller the σ value, the steeper the Gaussian curve, and more image details can be retained.

[0080] Next, the denoised image G1 is subjected to image correction to correct the distortion in the image, and the obtained image is recorded as G2. Image correction can include geometric correction and grayscale correction. Taking the perspective transformation method in geometric correction as an example, it projects the image to a new plane, mainly mapping from the original two-dimensional plane (x, y) to the new two-dimensional plane (x', y').

[0081] 102. Determine a palm print region image and a first palm print curve image of a corresponding palm print according to each palm image sample.

[0082] The palm print area image is an image of the area where the palm print is located. The first palm print curve image is a geometric curve image corresponding to the palm print texture, and is a binary image.

[0083] In some embodiments, see Figure 2 , Figure 2 1 is another flow chart of the palmprint recognition method provided by the embodiment of the present application. The above step 102 specifically includes the following steps 1021 and 1022, wherein:

[0084] 1021. Extract the image of the area where the corresponding palm print is located from each palm image sample to obtain a palm print area image.

[0085] Specifically, the above step 1021 may include:

[0086] Perform key point detection on each palm image sample to obtain multiple key points;

[0087] According to the image coordinates of the key point, the image of the area where the corresponding palm print is located is extracted from the palm image sample.

[0088] Among them, since there may be areas unrelated to palm prints in the original collected palm image samples, this area can seriously affect the extraction and recognition of palm print features. Therefore, in order to accurately locate the subsequent palm features, the pre-processed palm image sample G2 can be subjected to key point detection. These key points help define the geometric structure and features of the palm prints, eliminate non-palm print areas, and thus improve the accuracy of subsequent palm print recognition. Through key point detection, the palm image samples are standardized and aligned, which can reduce the recognition error caused by changes in palm position and angle. The n palm feature points at the valley between the fingers can be selected as key points, and the deep learning technology can be combined to detect and determine the coordinates of the palm feature points in each image sample G2, which are represented by the set β, β={(a,b),...,(a n ,b n )}, a and b are the coordinate components of the palm feature points. Afterwards, the position of the palm print area, that is, the position of the palm region of interest (ROI area), can be determined according to the coordinate β of the detected palm feature points, and the image of this area is extracted from the palm image sample as the palm print area image, denoted as G'. Usually, this area contains information such as the main line, wrinkles, fine textures, ridge ends, and bifurcation points of the palm print.

[0089] 1022. Input the palm print area image into a trained geometric curve generation model to generate a binary geometric curve image, and obtain a first palm print curve image corresponding to the palm print.

[0090] Among them, since the main lines and fine lines of the palm print play an important role in palm print recognition, and the muscle color and distribution have almost no effect on palm print recognition, in order to enhance the effective feature information such as the main lines, wrinkles, and bifurcation points of the palm print, the corresponding geometric curve image (i.e., the first palm print curve image S, see for details) can be drawn according to the palm print features in the palm print area image G'. Figure 3 ) to simulate the palm print texture, thereby improving the utilization rate of key feature information in the palm print area image. The drawing of the geometric curve image can be achieved by training a geometric curve generation model in advance, or other methods can be used, which are not limited here. When the drawing of the geometric curve image is achieved by training a geometric curve generation model in advance, the geometric curve generation model can be some models that can realize image conversion between different types (this solution mainly realizes the conversion from the real image collected by the camera to the drawn geometric curve image), such as BicycleGAN, CycleGAN, etc., but not limited to these models.

[0091] In some embodiments, the geometric curve generation model can be trained using a generative adversarial approach, and the training data includes real images and drawn geometric curve images. The real image can be the palm print area image of the known user mentioned above, or the palm print area image of other known users, which is not limited here. For example, when the real image can be the palm print area image of the known user mentioned above, the training process of the geometric curve model can include:

[0092] Generate a binary geometric curve image corresponding to each palmprint region image using the constructed geometric curve generation model to obtain a second palmprint curve image;

[0093] Generative adversarial training is performed on the geometric curve generation model according to the palmprint area image and the second palmprint curve image.

[0094] In the model training stage, the real palm print area image G' is first input into the geometric curve generation model to randomly generate a binary geometric curve image (i.e., the second palm print curve image), and then the real palm print area image G' and the second palm print curve image are simultaneously input into the geometric curve generation model as a pair of samples to perform generative adversarial training on the model. After the geometric curve generation model is trained, in the subsequent process of training the feature extraction model, the real palm print area image G' only needs to be input into the trained geometric curve generation model to obtain the first palm print curve image S with a high matching degree.

[0095] 103. According to the palmprint area image and the first palmprint curve image, a feature extraction model is trained to obtain a palmprint recognition model. The feature extraction model includes a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters. The first feature extraction module is used to extract features of the palmprint area image, and the second feature extraction module is used to extract features of the first palmprint curve image.

[0096] Among them, the first feature extraction module is mainly used to extract the overall features of the palmprint, and the second feature extraction module is mainly used to extract the key features of the main lines and fine lines of the palmprint. The two feature extraction modules share parameters during the training process. Parameter sharing means that different parts or all of the two networks share the same weights and bias parameters, rather than learning parameters independently. This can significantly reduce the number of parameters of the entire system and improve computational efficiency. Through parameter sharing, the two networks can jointly learn a set of parameters that perform better on a wider range of data, thereby improving the generalization ability of the model.

[0097] In some embodiments, please see Figure 2 , the above step 103 specifically includes:

[0098] 1031. Perform feature extraction on the palmprint region image using a first feature extraction module to obtain a first extracted feature, and perform feature extraction on the first palmprint curve image using a second feature extraction module to obtain a second extracted feature;

[0099] 1032. The feature extraction model is trained according to a target loss function, the first extracted feature, and the second extracted feature. The target loss function includes a weighted first loss function and a second loss function. The first loss function is used to determine a loss value according to features in a palmprint area image. The second loss function is used to determine a loss value according to features in the palmprint area image and features in the first palmprint curve image.

[0100] Furthermore, the above step 1032 specifically includes:

[0101] Determine a first loss value according to the first extracted feature and the first loss function, and determine a second loss value according to the first extracted feature, the second extracted feature and the second loss function;

[0102] Determine a comprehensive loss value by weighting according to the first loss value and the second loss value;

[0103] The parameters of the feature extraction model are adjusted according to the comprehensive loss value to train the feature extraction model.

[0104] Among them, the first loss function mainly helps the model focus on the overall features of the palmprint, and the second loss function mainly helps the model focus on the features of the main lines and fine lines of the palmprint. The weights of the first loss function and the second loss function can be adjusted as needed, so that the model pays more attention to samples that are helpful to improve the model performance during the training process, which helps the model learn more generalized features, thereby improving the performance of the model on unseen data and improving the palmprint recognition accuracy of the model.

[0105] See also Figure 4 , Figure 4 It is a schematic diagram of the data enhancement training process of the feature extraction model provided by the embodiment of the present application. Since the features of the palm print area image G' and the first palm print curve image S have similar statistical characteristics, the complexity of the model can be effectively reduced by sharing the parameters of the first feature extraction module and the second feature extraction module, and the training speed and generalization ability of the model can be improved. In addition to the first feature extraction module and the second feature extraction module, the feature extraction model also includes other network layers, such as an output layer for fusing and outputting the features extracted by the first feature extraction module (i.e., feature matrix) and the features extracted by the second feature extraction module (i.e., feature matrix), which is not limited here.

[0106] In the training process of the feature extraction model, the losses of G' and S should be considered at the same time, and the designed target loss function is: Loss Total = Loss1 + μ· Loss2, u is the weighting coefficient. Loss1 is the first loss function, Loss2 is the second loss function, Loss1 and Loss2 are calculated for the palmprint region image G', and Loss2 is calculated for the palmprint curve image. Loss1 is the main loss function, and Loss2 helps the model pay more attention to the main line and fine line features of the palmprint. Finally, the comprehensive loss Loss is calculated by weighted method. Total , and iteratively update the model parameters according to the loss value. Combined with the comprehensive loss Loss Total , extract key feature information such as the main lines and fine lines of the palmprint, and enhance the feature information of the main lines and fine lines of the palmprint, which greatly improves the utilization rate of palmprint information, and then realizes the palmprint data enhancement training of the model, which helps to improve the palmprint recognition accuracy of the model, and solves the current technical problem that it is difficult to support large-scale palmprint recognition tasks due to the small size of the palmprint data set, and makes a contribution to large-scale palmprint recognition tasks.

[0107] 104. Perform palmprint recognition based on the palmprint recognition model.

[0108] Among them, after the palmprint recognition model is obtained by training based on palm image samples of known users, the palmprint recognition model can be used to perform large-scale palmprint recognition tasks with a high recognition accuracy.

[0109] In some embodiments, see Figure 2 , the above step 104 specifically includes:

[0110] 1041. Acquire a target palm image of a user to be identified;

[0111] 1042. Determine a target palm print region image and a target palm print curve image of a corresponding palm print according to the target palm image;

[0112] 1043. Input the target palmprint curve image and the target palmprint area image into the palmprint recognition model to obtain palmprint feature data of the user to be recognized.

[0113] Among them, in the inference stage of the palmprint recognition model, the target palm image of the user to be identified can be first collected and preprocessed, key point detection can be performed on the preprocessed image, and the target palmprint area image can be extracted based on the detected key points. Then, the target palmprint area image can be input into the above-mentioned geometric curve generation model to obtain the target palmprint curve image. Then, the target palmprint curve image and the target palmprint area image can be simultaneously input into the above-mentioned palmprint recognition model to obtain palmprint feature data. Then, the palmprint feature data can be matched with the data in the palmprint feature library to determine the identity of the user to be identified. The palmprint feature library stores the palmprint feature data of known users.

[0114] As can be seen from the above, the palmprint recognition method provided by the embodiment of the present application obtains at least one palm image sample of each known user among multiple known users; determines the palmprint area image and the first palmprint curve image of the corresponding palmprint according to each palm image sample; trains the feature extraction model according to the palmprint area image and the first palmprint curve image to obtain a palmprint recognition model, the feature extraction model includes a first feature extraction module and a second feature extraction module with the same network structure and shared parameters, the first feature extraction module is used to extract the features of the palmprint area image, and the second feature extraction module is used to extract the features of the first palmprint curve image; palmprint recognition is performed based on the palmprint recognition model, that is, the model is trained using the image of the area where the palmprint is located and the curve image corresponding to the palmprint texture, so as to realize data enhancement training of the palmprint recognition model, ensure that the model can better learn the features of the main line and fine lines of the palmprint, so that the model can be well learned and generalized, thereby improving the recognition accuracy of the model for large-scale palmprint recognition tasks, and the palmprint recognition effect is good.

[0115] According to the method described in the above embodiment, the present application embodiment also provides a palmprint recognition device for executing the steps in the above palmprint recognition method. Figure 5 , Figure 5 2 is a schematic diagram of the structure of a palmprint recognition device 200 provided in an embodiment of the present application. The palmprint recognition device 200 is applied to an electronic device, and the palmprint recognition method can be executed by the electronic device or a functional module or functional entity in the electronic device. Specifically, the palmprint recognition device 200 includes an acquisition unit 201, a determination unit 202, a training unit 203 and a recognition unit 204, wherein:

[0116] An acquisition unit 201 is used to acquire at least one palm image sample of each known user among a plurality of known users;

[0117] A determination unit 202, configured to determine a palm print region image and a first palm print curve image of a corresponding palm print according to each palm image sample;

[0118] The training unit 203 is used to train a feature extraction model according to the palmprint region image and the first palmprint curve image to obtain a palmprint recognition model, wherein the feature extraction model includes a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is used to extract features of the palmprint region image, and the second feature extraction module is used to extract features of the first palmprint curve image;

[0119] The recognition unit 204 is used to perform palmprint recognition based on the palmprint recognition model.

[0120] In some embodiments, the determining unit 202 is specifically configured to:

[0121] Extracting an image of the area where the corresponding palm print is located from each palm image sample to obtain a palm print area image;

[0122] The palmprint region image is input into a trained geometric curve generation model to generate a binary geometric curve image, thereby obtaining a first palmprint curve image corresponding to the palmprint.

[0123] In some embodiments, the determining unit 202 is specifically configured to:

[0124] Perform key point detection on each palm image sample to obtain multiple key points;

[0125] According to the image coordinates of the key point, an image of the area where the corresponding palm print is located is extracted from the palm image sample.

[0126] In some embodiments, the training unit 203 is further configured to:

[0127] Generate a binary geometric curve image corresponding to each palmprint region image using the constructed geometric curve generation model to obtain a second palmprint curve image;

[0128] Generative adversarial training is performed on the geometric curve generation model according to the palmprint area image and the second palmprint curve image.

[0129] In some embodiments, the training unit 203 is specifically used to:

[0130] Using the first feature extraction module to perform feature extraction on the palmprint region image to obtain a first extracted feature, and using the second feature extraction module to perform feature extraction on the first palmprint curve image to obtain a second extracted feature;

[0131] The feature extraction model is trained according to the target loss function, the first extracted feature and the second extracted feature. The target loss function includes a weighted first loss function and a second loss function. The first loss function is used to determine the loss value according to the features in the palmprint area image, and the second loss function is used to determine the loss value according to the features in the palmprint area image and the features in the first palmprint curve image.

[0132] In some embodiments, the training unit 203 is specifically used to:

[0133] Determine a first loss value according to the first extracted feature and the first loss function, and determine a second loss value according to the first extracted feature, the second extracted feature and the second loss function;

[0134] Determine a comprehensive loss value by weighting according to the first loss value and the second loss value;

[0135] The parameters of the feature extraction model are adjusted according to the comprehensive loss value to train the feature extraction model.

[0136] In some embodiments, the identification unit 204 is specifically used to:

[0137] Obtain a target palm image of a user to be identified;

[0138] Determine a target palm print region image and a target palm print curve image of the corresponding palm print according to the target palm image;

[0139] The target palmprint curve image and the target palmprint area image are input into the palmprint recognition model to obtain palmprint feature data of the user to be recognized.

[0140] It should be noted that the specific details of each module unit in the palmprint recognition device 200 have been described in detail in the embodiment of the palmprint recognition method, and will not be repeated here.

[0141] As can be seen from the above, the palmprint recognition device 200 provided in the embodiment of the present application obtains at least one palm image sample of each known user among multiple known users through the acquisition unit 201; the determination unit 202 determines the palmprint area image and the first palmprint curve image of the corresponding palmprint according to each palm image sample; the training unit 203 trains the feature extraction model according to the palmprint area image and the first palmprint curve image to obtain a palmprint recognition model, the feature extraction model includes a first feature extraction module and a second feature extraction module with the same network structure and shared parameters, the first feature extraction module is used to extract the features of the palmprint area image, and the second feature extraction module is used to extract the features of the first palmprint curve image; the recognition unit 204 performs palmprint recognition based on the palmprint recognition model, that is, uses the image of the palmprint area and the curve image corresponding to the palmprint texture to train the model, so as to realize data enhancement training of the palmprint recognition model, ensure that the model can be well learned and generalized, thereby improving the recognition accuracy of the model for large-scale palmprint recognition tasks, and the palmprint recognition effect is good.

[0142] In some embodiments, the palmprint recognition device in the embodiments of the present application can be an electronic device, or a component in an electronic device, such as an integrated circuit or a chip. The electronic device can be a server, which can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. The embodiments of the present application are not specifically limited.

[0143] In some embodiments, Figure 6 As shown, the embodiment of the present application also provides an electronic device 300, including a processor 301, a memory 302, and a computer program stored in the memory 302 and executable on the processor 301. When the program is executed by the processor 301, each process of the above-mentioned palmprint recognition method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0144] It should be noted that the electronic devices in the embodiments of the present application include the above-mentioned mobile electronic devices and non-mobile electronic devices.

[0145] Figure 7 A schematic diagram of the hardware structure of an electronic device to implement an embodiment of the present application.

[0146] The electronic device 400 includes but is not limited to: a radio frequency unit 401, a network module 402, an audio output unit 403, an input unit 404, a sensor 405, a display unit 406, a user input unit 407, an interface unit 408, a memory 409, a processor 410 and other components.

[0147] In some embodiments, the processor 410 implements the following steps when executing the program:

[0148] obtaining at least one palm image sample of each known user among a plurality of known users;

[0149] Determine a palmprint region image and a first palmprint curve image of a corresponding palmprint according to each palm image sample;

[0150] According to the palmprint region image and the first palmprint curve image, a feature extraction model is trained to obtain a palmprint recognition model, wherein the feature extraction model includes a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is used to extract features of the palmprint region image, and the second feature extraction module is used to extract features of the first palmprint curve image;

[0151] Palmprint recognition is performed based on the palmprint recognition model.

[0152] In some embodiments, the processor 410 is specifically configured to:

[0153] Extracting an image of the area where the corresponding palm print is located from each palm image sample to obtain a palm print area image;

[0154] The palmprint region image is input into a trained geometric curve generation model to generate a binary geometric curve image, thereby obtaining a first palmprint curve image corresponding to the palmprint.

[0155] In some embodiments, the processor 410 is further configured to:

[0156] Generate a binary geometric curve image corresponding to each palmprint region image using the constructed geometric curve generation model to obtain a second palmprint curve image;

[0157] Generative adversarial training is performed on the geometric curve generation model according to the palmprint area image and the second palmprint curve image.

[0158] In some embodiments, the processor 410 is specifically configured to:

[0159] Perform key point detection on each palm image sample to obtain multiple key points;

[0160] According to the image coordinates of the key point, an image of the area where the corresponding palm print is located is extracted from the palm image sample.

[0161] In some embodiments, the processor 410 is specifically configured to:

[0162] Using the first feature extraction module to perform feature extraction on the palmprint region image to obtain a first extracted feature, and using the second feature extraction module to perform feature extraction on the first palmprint curve image to obtain a second extracted feature;

[0163] The feature extraction model is trained according to the target loss function, the first extracted feature and the second extracted feature. The target loss function includes a weighted first loss function and a second loss function. The first loss function is used to determine the loss value according to the features in the palmprint area image, and the second loss function is used to determine the loss value according to the features in the palmprint area image and the features in the first palmprint curve image.

[0164] In some embodiments, the processor 410 is specifically configured to:

[0165] Determine a first loss value according to the first extracted feature and the first loss function, and determine a second loss value according to the first extracted feature, the second extracted feature and the second loss function;

[0166] Determine a comprehensive loss value by weighting according to the first loss value and the second loss value;

[0167] The parameters of the feature extraction model are adjusted according to the comprehensive loss value to train the feature extraction model.

[0168] In some embodiments, the processor 410 is specifically configured to:

[0169] Obtain a target palm image of a user to be identified;

[0170] Determine a target palm print region image and a target palm print curve image of the corresponding palm print according to the target palm image;

[0171] The target palmprint curve image and the target palmprint area image are input into the palmprint recognition model to obtain palmprint feature data of the user to be recognized.

[0172] Those skilled in the art will appreciate that the electronic device 400 may also include a power source (such as a battery) for supplying power to each component, and the power source may be logically connected to the processor 410 through a power management system, thereby implementing functions such as managing charging, discharging, and power consumption management through the power management system. Figure 6 The electronic device structure shown in the figure does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently, which will not be described in detail here.

[0173] It should be understood that in the embodiment of the present application, the input unit 404 may include a graphics processor (Graphics Processing Unit, GPU) 4041 and a microphone 4042, and the graphics processor 4041 processes the image data of the static picture or video obtained by the image capture device (such as a camera) in the video capture mode or the image capture mode. The display unit 406 may include a display panel 4061, and the display panel 4061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 407 includes a touch panel 4071 and at least one of other input devices 4072. The touch panel 4071 is also called a touch screen. The touch panel 4071 may include two parts: a touch detection device and a touch controller. Other input devices 4072 may include, but are not limited to, a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.

[0174] The memory 409 can be used to store software programs and various data. The memory 409 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instructions required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 409 may include a volatile memory or a non-volatile memory, or the memory 409 may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM). The memory 409 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.

[0175] The processor 410 may include one or more processing units; the processor 410 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 410.

[0176] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the above-mentioned palmprint recognition method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0177] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0178] The embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned palmprint recognition method when executed by a processor.

[0179] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0180] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0181] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0182] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

[0183] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0184] In the description of the present application, “plurality” means two or more.

[0185] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0186] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A palmprint recognition method, characterized in that: include: obtaining at least one palm image sample of each known user among a plurality of known users; Determine a palmprint region image and a first palmprint curve image of a corresponding palmprint according to each of the palm image samples; According to the palmprint area image and the first palmprint curve image, a feature extraction model is trained to obtain a palmprint recognition model, wherein the feature extraction model includes a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is used to extract features of the palmprint area image, and the second feature extraction module is used to extract features of the first palmprint curve image; Palmprint recognition is performed based on the palmprint recognition model.

2. The palmprint recognition method according to claim 1, characterized in that: The step of determining a palm print region image and a first palm print curve image of a corresponding palm print according to each of the palm image samples comprises: Extracting an image of the area where the corresponding palm print is located from each of the palm image samples to obtain a palm print area image; The palmprint region image is input into a trained geometric curve generation model to generate a binary geometric curve image, thereby obtaining a first palmprint curve image corresponding to the palmprint.

3. The palmprint recognition method according to claim 2, characterized in that: The training process of the geometric curve generation model includes: Generate a binary geometric curve image corresponding to each of the palmprint area images using the constructed geometric curve generation model to obtain a second palmprint curve image; Generative adversarial training is performed on the geometric curve generation model according to the palmprint area image and the second palmprint curve image.

4. The palmprint recognition method according to claim 2, characterized in that: The step of extracting an image of an area where a corresponding palm print is located from each of the palm image samples comprises: Performing key point detection on each palm image sample to obtain multiple key points; According to the image coordinates of the key points, an image of the area where the corresponding palm print is located is extracted from the palm image sample.

5. The palmprint recognition method according to claim 1, characterized in that: The step of training a feature extraction model according to the palmprint region image and the first palmprint curve image comprises: Using the first feature extraction module to perform feature extraction on the palmprint region image to obtain a first extracted feature, and using the second feature extraction module to perform feature extraction on the first palmprint curve image to obtain a second extracted feature; The feature extraction model is trained according to a target loss function, the first extracted features and the second extracted features, wherein the target loss function includes a weighted first loss function and a second loss function, wherein the first loss function is used to determine a loss value according to features in the palmprint area image, and the second loss function is used to determine a loss value according to features in the palmprint area image and features in the first palmprint curve image.

6. The palmprint recognition method according to claim 5, characterized in that: The step of training the feature extraction model according to the target loss function, the first extracted features, and the second extracted features includes: Determine a first loss value according to the first extracted feature and the first loss function, and determine a second loss value according to the first extracted feature, the second extracted feature, and the second loss function; Determining a comprehensive loss value by weighting according to the first loss value and the second loss value; The parameters of the feature extraction model are adjusted according to the comprehensive loss value to train the feature extraction model.

7. The palmprint recognition method according to any one of claims 1 to 6, characterized in that: The performing palmprint recognition based on the palmprint recognition model comprises: Obtain a target palm image of a user to be identified; Determine a target palm print region image and a target palm print curve image of a corresponding palm print according to the target palm image; The target palmprint curve image and the target palmprint area image are input into the palmprint recognition model to obtain palmprint feature data of the user to be recognized.

8. A palmprint recognition device, characterized in that: include: an acquisition unit, configured to acquire at least one palm image sample of each known user among a plurality of known users; A determination unit, configured to determine a palm print region image and a first palm print curve image of a corresponding palm print according to each of the palm image samples; a training unit, configured to train a feature extraction model according to the palmprint region image and the first palmprint curve image to obtain a palmprint recognition model, wherein the feature extraction model comprises a first feature extraction module and a second feature extraction module having the same network structure and sharing parameters, wherein the first feature extraction module is configured to extract features of the palmprint region image, and the second feature extraction module is configured to extract features of the first palmprint curve image; The recognition unit is used to perform palmprint recognition based on the palmprint recognition model.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the palmprint recognition method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the palmprint recognition method according to any one of claims 1 to 7 is implemented.

Images