Identity verification method and apparatus, device and medium

Abstract

An identity verification method is executed by a computer device. The method comprises: in response to an identity verification trigger event, acquiring at least one palm image collected at a preset collection field of view (202); detecting key points in each palm image to obtain respective palm key points of each palm image (204); according to the respective palm key points of each palm image, determining the palm direction of a palm in each palm image (206); for each palm image, determining a preset calibration direction in the collection field of view, and determining the collection angle of the palm in the palm image according to an included angle between the palm direction of the palm in the palm image and the calibration direction (208); screening from the at least one palm image a palm image having an acquisition angle within a preset collection angle range, so as to obtain a target palm image (210); and performing identity verification according to the target palm image (212).

Description

Authentication methods, devices, equipment, and media

[0001] Related applications

[0002] This application claims priority to Chinese patent application number 2023107544814, filed on June 25, 2023, entitled “Resource Transfer Method, Apparatus, Equipment and Medium,” the entire text of which is incorporated herein by reference. Technical Field

[0003] The present application relates to computer technology, and in particular to an identity authentication method, apparatus, device and medium. Background Art

[0004] With the development of computer technology, palm verification payment technology has emerged. With the widespread application of palm verification payment in daily life, it can provide people with more choices and convenience in payment methods, and people can make secure payments by palm verification.

[0005] In the palm verification payment scenario, the environment is usually complex and there are cases of misjudgment during identity verification.

[0006] For example, in offline palm verification payment scenarios such as convenience stores and supermarkets, if the user has no intention to pay, an erroneous payment may occur because the user mistakenly places his palm in the collection field of view of the palm verification device.

[0007] Summary of the Invention

[0008] According to various embodiments of the present application, an identity authentication method, apparatus, device, and medium are provided.

[0009] In a first aspect, the present application provides an identity authentication method, the method comprising:

[0010] In response to an identity authentication triggering event, acquiring at least one palm image captured with a preset capture field of view;

[0011] Detecting key points in each of the palm images respectively to obtain palm key points of each of the palm images;

[0012] Determining the palm direction of each palm image according to the palm key points of each palm image;

[0013] For each of the palm images, determining a preset calibration direction in the acquisition field of view, and determining an acquisition angle of the palm in the palm image according to an angle between the palm direction of the palm in the palm image and the calibration direction;

[0014] Filtering palm images whose acquisition angles are within a preset acquisition angle range from the at least one palm image to obtain a target palm image;

[0015] Authentication is performed based on the target palm image.

[0016] In a second aspect, the present application provides an identity authentication device, the device comprising:

[0017] an acquisition module, configured to acquire at least one palm image captured with a preset capture field of view in response to an identity authentication triggering event;

[0018] A detection module, configured to detect key points in each of the palm images to obtain palm key points of each of the palm images;

[0019] a determination module, configured to determine the palm direction of the palm in each of the palm images based on the respective palm key points of each of the palm images; determine a preset calibration direction in the acquisition field of view for each of the palm images, and determine the acquisition angle of the palm in the palm image based on the angle between the palm direction of the palm in the palm image and the calibration direction;

[0020] A screening module, configured to screen palm images whose acquisition angles are within a preset acquisition angle range from the at least one palm image to obtain a target palm image;

[0021] The identity authentication module is used to perform identity authentication based on the target palm image.

[0022] In a third aspect, the present application provides a computer device including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps in the method embodiments of the present application are implemented.

[0023] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the various method embodiments of the present application.

[0024] In a fifth aspect, the present application provides a computer program product, including a computer program, which implements the steps in the method embodiments of the present application when the computer program is executed by a processor.

[0025] The details of one or more embodiments of the present application are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the present application will become apparent from the description, drawings, and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the conventional technology, the following briefly introduces the drawings required for use in the embodiments or the conventional technology descriptions. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the disclosed drawings without any creative work.

[0027] FIG1 is a diagram illustrating an application environment of an identity authentication method according to an embodiment;

[0028] FIG2 is a schematic diagram of a process flow of an identity authentication method according to an embodiment;

[0029] FIG3 is a schematic diagram of palm key points detected from a palm image in one embodiment;

[0030] FIG4 is a schematic diagram of verifying an erroneous payment by the initiator in one embodiment;

[0031] FIG5 is a schematic diagram of screening a target palm image from various palm images in one embodiment;

[0032] FIG6 is a schematic diagram of a palm image acquisition and payment process in one embodiment;

[0033] FIG7 is a schematic diagram of a process for defining a calibration direction in one embodiment;

[0034] FIG8 is a schematic diagram of palm direction and calibration direction collected for the left hand in one embodiment;

[0035] FIG9 is a schematic diagram of a calculation process of a collection angle for left-hand collection in one embodiment;

[0036] FIG10 is a schematic diagram of a calculation process of a collection angle for left-hand collection in another embodiment;

[0037] FIG11 is a schematic diagram of palm direction and calibration direction collected for the right hand in one embodiment;

[0038] FIG12 is a schematic diagram of a calculation process of the acquisition angle for right-hand acquisition in one embodiment;

[0039] FIG13 is a schematic diagram of a palm area block diagram in one embodiment;

[0040] FIG14 is a schematic diagram of a palm image obtained based on a palm region frame in one embodiment;

[0041] FIG15 is a schematic diagram of palm key points detected from a captured palm image in one embodiment;

[0042] FIG16 is a schematic flow chart of an identity authentication method in another embodiment;

[0043] FIG17 is a block diagram of a structure of an identity verification device according to an embodiment;

[0044] FIG18 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0046] The identity authentication method provided in this application can be applied in the application environment shown in Figure 1. Terminal 102 communicates with server 104 via a network. A data storage system can be separately provided to store data that server 104 needs to process. The data storage system can be integrated with server 104 or placed in the cloud or on another server. Terminal 102 can be, but is not limited to, a palm authentication device, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart car devices, etc. Portable wearable devices can include smart watches, smart bracelets, head-mounted devices, etc. Server 104 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides network security services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, cloud security, host security, CDN, and basic cloud computing services such as big data and artificial intelligence platforms. Terminal 102 and server 104 can be connected directly or indirectly via wired or wireless communication, which is not limited in this application.

[0047] In response to an identity authentication trigger event, terminal 102 may obtain at least one palm image captured within a preset capture field of view. For each palm image, terminal 102 may perform key point detection on the palm image to obtain palm key points. Terminal 102 may determine the palm orientation of the palm in each palm image based on the palm key points of each palm image. For each palm image, terminal 102 may also determine a preset calibration direction within the capture field of view, and determine the capture angle of the palm in the palm image based on the angle between the palm orientation of the palm in the palm image and the calibration direction. Furthermore, terminal 102 may filter palm images from the at least one palm image for those with capture angles within the preset capture angle range to obtain a target palm image, and perform identity authentication based on the target palm image.

[0048] It is understood that the terminal 102 can capture palm images using an image capture unit deployed on the terminal. It is also understood that the server 104 stores palm images previously captured by the image capture unit, and the terminal 102 can also obtain the palm images from the server. This embodiment is not limited to this. It is understood that the application scenario in Figure 1 is merely illustrative and is not limited to this.

[0049] It should be noted that the identity authentication methods in some embodiments of the present application use artificial intelligence technology. For example, the key points of the palm in the present application are detected using artificial intelligence technology. In order to make it easier to understand artificial intelligence, the concept of artificial intelligence is now explained. Specifically, artificial intelligence is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that the machines have the functions of perception, reasoning and decision-making. The present application implements key point detection for palm images based on artificial intelligence technology, which can improve the accuracy of key point detection.

[0050] In one embodiment, as shown in FIG2 , an identity authentication method is provided. This embodiment is described by taking the method as applied to the terminal 102 in FIG1 as an example. The terminal 102 may be a palm authentication device. The method includes the following steps:

[0051] Step 202: In response to an identity authentication triggering event, obtaining at least one palm image captured with a preset capture field of view.

[0052] Among them, the identity authentication trigger event is an event used to trigger an identity authentication action, which can be a payment trigger operation, or an event in which a physical palm is detected within the capture field of view, or the recognition of other scenarios that require identity authentication. The palm image is captured by the image acquisition unit of the terminal, and the capture field of view is the range in which the terminal captures the palm image from the environment through its image acquisition unit. For a terminal, the capture field of view of its image acquisition unit is also fixed relative to the terminal. The terminal can be a fixed terminal or a mobile terminal. The image acquisition unit is a hardware unit that can capture images, which can be called a camera. The palm image is an image that includes the image of the palm. The palm in the palm image refers to the image of the palm formed by the physical palm being imaged in the palm image.

[0053] In one embodiment, the image acquisition unit is deployed on the terminal. When the user places the palm above the image acquisition unit, the terminal can capture an image of the palm through the image acquisition unit and directly use the captured image as the palm image.

[0054] In one embodiment, the terminal may be a palm verification device, and the image acquisition unit may be a camera. When the user places the palm of the user in the acquisition field of view of the camera of the palm verification device, the terminal may capture the palm image through the camera to obtain the palm image.

[0055] Step 204 : Detect key points in each palm image to obtain palm key points of each palm image.

[0056] Key points are characteristic points that mark the location of the palm in a palm image. Detecting at least two key points allows the palm to be located in the palm image. Palm key points are key points detected from a palm image. They can be located in the palm or on the edges of the palm, specifically between fingers, at the fingertips, or in the palm.

[0057] In some embodiments, the position of the palm keypoint in the palm image is fixed relative to the palm itself. When the palm position changes, the palm keypoint also changes with the palm position. If the palm keypoint is a finger gap point, the position of the palm keypoint does not change when the palm is open or contracted.

[0058] In one embodiment, the number of detected palm key points may be at least two. As shown in FIG3 , after acquiring the palm image captured by the image acquisition unit, the terminal may perform key point detection on the palm image to obtain palm key points of the palm in the palm image, namely, palm key point 1, palm key point 2, palm key point 3, and palm key point 4.

[0059] In one embodiment, for each palm image, the terminal may input the palm image into a trained palm key point detection model to perform key point detection on the palm image through the palm key point detection model and output the palm key points of the palm in the palm image.

[0060] In one embodiment, a trained palm key point detection model can be obtained by training in the following manner: a terminal can obtain a training image including a palm, wherein the training image carries reference palm key points calibrated for the palm. The terminal can perform key point detection on the training image using the palm key point detection model to be trained to obtain predicted palm key points, and determine a loss value based on the positional differences between the predicted palm key points and the reference palm key points. The terminal can then use this loss value to train the palm key point detection model to be trained, thereby obtaining a trained palm key point detection model.

[0061] Step 206: Determine the palm direction of each palm image based on the palm key points of each palm image.

[0062] The palm direction represents the orientation of the palm in the palm image. Palm direction can be represented in different ways, such as the direction from the palm to a specific fingertip, or the direction formed by two finger joints. These two directions are different, but both represent palm direction.

[0063] Specifically, the number of the detected palm key points may include at least two. The terminal may determine the direction of the palm in the palm image according to the at least two detected palm key points to obtain the palm direction.

[0064] In one embodiment, the terminal may determine the palm shape according to the detected at least two palm key points, and thereby determine the palm direction from the palm shape according to a predefined definition of the palm direction.

[0065] In one embodiment, the terminal may construct a vector based on the at least two detected palm key points, and use the direction of the constructed vector as the direction of the palm in the palm image to obtain the palm direction.

[0066] In one embodiment, the terminal may randomly select two palm key points from the detected palm key points and construct a vector based on the two selected palm key points. The terminal may then use the direction of the constructed vector as the direction of the palm in the palm image to obtain the palm direction.

[0067] In one embodiment, the terminal may randomly select two palm key points from the detected palm key points and use the two selected palm key points as the starting point and ending point of a vector to construct a vector. Furthermore, the terminal may use the direction of the constructed vector as the direction of the palm in the palm image to obtain the palm direction.

[0068] Step 208 : for each palm image, determine a preset calibration direction in the acquisition field of view, and determine the acquisition angle of the palm in the palm image according to the angle between the palm direction of the palm in the palm image and the calibration direction.

[0069] The calibration direction is a direction predefined for the image acquisition unit, and the acquisition angle is the angle between the palm direction of the palm in the palm image and the calibration direction of the image acquisition unit.

[0070] In one embodiment, the terminal may determine a direction angle between the palm direction and the calibration direction according to the difference between the palm direction and the calibration direction of the image acquisition unit, and use the direction angle as the acquisition angle of the palm in the palm image.

[0071] In one embodiment, the terminal may construct a vector along the palm direction and a vector along the calibration direction of the image acquisition unit. Furthermore, the terminal may determine the angle between the palm direction and the calibration direction based on the vector constructed along the palm direction and the vector constructed along the calibration direction, and use the angle as the palm acquisition angle in the palm image.

[0072] Step 210 , screening palm images whose acquisition angles are within a preset acquisition angle range from at least one palm image to obtain a target palm image.

[0073] In one embodiment, the terminal may select, from the at least one captured palm image, palm images whose capture angles fall within a preset capture angle range. It is understood that the number of selected palm images may be at least one. The terminal may then randomly select one palm image from the selected palm images as a target palm image.

[0074] In one embodiment, the terminal may preliminarily screen at least one captured palm image for palm images whose capture angles fall within a preset capture angle range. It is understood that the number of screened palm images may be at least one. The terminal may perform image quality testing on each of the preliminarily screened palm images and select the palm image with the best quality as the target palm image. It is understood that the image quality testing may include testing at least one of palm integrity and clarity.

[0075] In one embodiment, as shown in FIG4 , in a palm verification payment scenario, the terminal may be a palm verification device 401. When the verification initiator 402 (the payee, such as a cashier in a supermarket) uses his palm to signal the payee (such as a customer in a supermarket) to make a palm verification payment, the palm verification device may capture the palm image of the verification initiator 402, thereby causing the verification initiator to be mistakenly identified as the payee. The preset acquisition angle range of the present application can be defined as the range of acquisition angles corresponding to the palm image captured for the payee. In an embodiment of the present application, the target palm image captured for the payee is screened out from the captured palm images by the acquisition angle, so as to avoid the verification initiator from making an erroneous payment at the software level, thereby reducing the cost of using the palm verification device.

[0076] In one embodiment, in a palm verification payment scenario, the terminal serves as a palm verification device. As shown in FIG5 , the palm verification device may screen the captured palm images (i.e., the eight palm images 401-408) for palm images whose acquisition angles fall within a preset acquisition angle range (e.g., the acquisition angles corresponding to the five palm images 401-405 fall within the preset acquisition angle range of 0-90°). The palm verification device may then randomly select one of the five screened palm images as the target palm image (e.g., 403 is the target palm image).

[0077] In one embodiment, in a palm verification payment scenario, the terminal serves as a palm verification device. As shown in Figure 6, during the image capture phase, after the verification initiator (e.g., a supermarket cashier) initiates a payment request (e.g., creating an order), the palm verification device captures an image of the palm above the palm verification device, obtaining at least one palm image. The palm verification device then selects a target palm image from the at least one palm image.

[0078] Step 212: Perform identity authentication based on the target palm image.

[0079] After determining the target palm image, the terminal may obtain biometric information of the palm in the target palm image and perform identity verification based on the biometric information in the target palm image. If the identity verification fails, the payment operation may be abandoned. If the identity verification succeeds, the payment operation may be performed in response to the successful identity verification based on the biometric information in the target palm image. The biometric information may include at least one of palm prints, palm shape, or color distribution on the palm.

[0080] During identity authentication, the user can first determine whether the palm in the target palm image is registered in the database. If not, authentication fails; otherwise, authentication succeeds. Upon successful authentication, the user of the target palm image can be further determined, allowing payment operations to be performed using that user's account. Specifically, when performing a payment operation, order information can be obtained, the payment amount can be obtained from the order information, and the payment operation can be performed based on the payment amount.

[0081] In the above-mentioned identity verification method, in response to an identity verification trigger event, at least one palm image captured within a preset capture field of view is obtained, key points are detected in each palm image, and palm key points are obtained for each palm image. Based on the palm key points of each palm image, the palm orientation of the palm in each palm image is determined. For each palm image, a preset calibration direction within the capture field of view is determined, and the palm capture angle of the palm in the palm image is determined based on the angle between the palm orientation of the palm in the palm image and the calibration direction. This automatically filters palm images from the at least one palm image for those with capture angles within the preset capture angle range used to characterize the payee, obtaining a target palm image, and performing identity verification based on the target palm image. This method ensures that verification is only performed correctly when the palm is in a specific orientation, thus avoiding misjudgments.

[0082] When the above-mentioned identity authentication method is applied to the palm verification payment scenario, compared with the traditional method of adding a physical baffle next to the palm verification device to avoid the verification initiator's erroneous payment at the physical level, the present application detects the palm key points of the collected palm image, and automatically determines the palm collection angle based on the detected palm key points, and then filters out the target palm image for palm verification payment from the collected palm images based on the collection angle, so as to avoid the verification initiator's erroneous payment at the software level, thereby reducing the use cost of the palm verification device.

[0083] In one embodiment, the palm direction of the palm in each palm image is determined based on the palm key points of each palm image, including: for each palm image, if the palm key points of the palm image include a first palm key point and a second palm key point, constructing a first vector based on the first palm key point and the second palm key point, and using the direction of the first vector as the palm direction of the palm in the palm image.

[0084] The palm image has at least two palm key points, including at least a first palm key point and a second palm key point. The first vector is a vector constructed based on the first palm key point and the second palm key point.

[0085] Specifically, the terminal may use either the first palm key point or the second palm key point as the starting point of the first vector, and construct the first vector based on the first palm key point and the second palm key point. It is understood that the first vector passes through the first palm key point and the second palm key point. The terminal may use the direction of the first vector as the direction of the palm in the palm image to obtain the palm direction.

[0086] In this embodiment, the first palm key point and the second palm key point are used to construct a first vector, and the direction of the first vector is used as the palm direction. The calculation is simple and the palm direction can be determined efficiently, thereby achieving efficient identity authentication.

[0087] In one embodiment, the terminal may use any one of the first palm key point and the second palm key point as the starting point of the first vector, and use the key point of the first palm key point and the second palm key point that is not used as the starting point of the first vector as the ending point of the first vector to construct the first vector. It can be understood that the modulus of the first vector is the length of the line segment between the first palm key point and the second palm key point. For example, if the terminal uses the first palm key point as the starting point of the first vector, the second palm key point may be used as the ending point of the first vector to construct the first vector. If the terminal uses the second palm key point as the starting point of the first vector, the first palm key point may be used as the ending point of the first vector to construct the first vector.

[0088] In one embodiment, the terminal may use either the first palm key point or the second palm key point as the starting point of the first vector, and use the key point of the first palm key point or the second palm key point that is not the starting point of the first vector as the point through which the first vector passes, to construct the first vector. It is understood that the modulus of the first vector is greater than the length of the line segment between the first palm key point and the second palm key point.

[0089] In one embodiment, the terminal may use either the first palm key point or the second palm key point as the starting point of the first vector, and use the direction of the key point of the first palm key point or the second palm key point that is not the starting point of the first vector as the direction of the first vector to construct the first vector. It is understood that the first vector does not pass through the second palm key point, and the modulus of the first vector is less than the length of the line segment between the first palm key point and the second palm key point.

[0090] In one embodiment, for each palm image, a preset calibration direction in the acquisition field of view is determined respectively, and the acquisition angle of the palm in the palm image is determined based on the angle between the palm direction of the palm in the palm image and the calibration direction, including: for each palm image, if the palm key points of the palm image include a first palm key point and a second palm key point, using the first palm key point or the second palm key point as the starting point, constructing a second vector along the horizontal direction, and using the direction of the second vector as the preset calibration direction in the acquisition field of view.

[0091] The first palm key point or the second palm key point can be used as the starting point of the second vector. The terminal can use the starting point of the first vector as a new starting point, construct a second vector along the horizontal direction, and use the direction of the second vector as the preset calibration direction for the current palm image in the captured field of view.

[0092] In one embodiment, as shown in part (a) of Figure 7 , by placing the scissors 702 directly above the image acquisition unit of the terminal 701, the horizontal direction in this application can be defined as the direction shown in part (b) of Figure 7 . It is understood that the direction perpendicular to the horizontal direction is the vertical direction.

[0093] In the above embodiment, constructing a first vector using two palm key points and using the direction of the first vector as the palm direction of the palm in the palm image can improve the accuracy of palm direction acquisition. Furthermore, by using the starting point of the first vector as a new starting point, constructing a second vector along the horizontal direction and using the direction of the second vector as the calibration direction, the accuracy of calibration direction acquisition can be improved. Furthermore, determining the palm acquisition angle in the palm image based on the more accurate palm direction and calibration direction can further improve the accuracy of the acquisition angle and thus the accuracy of identity verification.

[0094] In one embodiment, constructing a second vector along the horizontal direction with the first palm key point or the second palm key point as the starting point includes: when the palm in the palm image is recognized as the left palm, constructing the second vector along the first horizontal direction with the first palm key point or the second palm key point as the starting point; when the palm in the palm image is recognized as the right palm, constructing the second vector along the second horizontal direction with the first palm key point or the second palm key point as the starting point; wherein the first horizontal direction and the second horizontal direction are opposite horizontal directions.

[0095] The horizontal direction includes a first horizontal direction and a second horizontal direction, and the first horizontal direction is opposite to the second horizontal direction. The left palm image is a palm image captured by the image acquisition unit for the user's left hand. The right palm image is a palm image captured by the image acquisition unit for the user's right hand.

[0096] In one embodiment, when the palm image is a left palm image, the terminal can use the first palm key point as a new starting point to construct a second vector along the first horizontal direction; when the palm image is a right palm image, the terminal can use the first palm key point as a new starting point to construct a second vector along the second horizontal direction.

[0097] In one embodiment, the terminal may use the first palm key point as the starting point of the first vector. In the case where the palm image is a left palm image obtained by capturing the user's left hand, the terminal may use the first palm key point (i.e., the starting point of the first vector) as the new starting point to construct a second vector along the first horizontal direction. In the case where the palm image is a right palm image obtained by capturing the user's right hand, the terminal may use the first palm key point (i.e., the starting point of the first vector) as the new starting point to construct a second vector along the second horizontal direction. It can be understood that the first vector and the second vector have a common starting point, i.e., the starting points of the first vector and the second vector are both the first palm key point.

[0098] In the above embodiment, when the palm image is captured for the user's left hand, the accuracy of constructing the second vector can be improved by constructing the second vector along the first horizontal direction using the first palm key point as the new starting point. When the palm image is captured for the user's right hand, the accuracy of constructing the second vector can be improved by constructing the second vector along the second horizontal direction using the first palm key point as the new starting point. This can improve the accuracy of constructing the second vector when the palm image is captured for the user's right hand. Thus, determining the calibration direction based on the more accurate second vector can improve the accuracy of the calibration direction.

[0099] In one embodiment, when the palm in the palm image is identified as the left palm, a second vector is constructed along the first horizontal direction with the first palm key point or the second palm key point as the starting point, including: when the palm in the palm image is identified as the left palm, a second vector is constructed along the positive direction of the horizontal axis of the image coordinate system of the palm image with the first palm key point as the starting point; when the palm in the palm image is identified as the right palm, a second vector is constructed along the second horizontal direction with the first palm key point or the second palm key point as the starting point, including: when the palm in the palm image is identified as the right palm, a second vector is constructed along the negative direction of the horizontal axis of the image coordinate system of the palm image with the first palm key point as the starting point.

[0100] The image coordinate system is a coordinate system established based on the palm image. The first horizontal direction includes the positive direction of the horizontal axis of the image coordinate system, and the second horizontal direction includes the negative direction of the horizontal axis of the image coordinate system.

[0101] Specifically, if the palm image is a left palm image obtained by capturing the user's left hand, the terminal can use the first palm key point (i.e., the starting point of the first vector) as a new starting point and construct a second vector along the positive direction of the horizontal axis of the image coordinate system of the left palm image. If the palm image is a right palm image obtained by capturing the user's right hand, the terminal can use the first palm key point (i.e., the starting point of the first vector) as a new starting point and construct a second vector along the negative direction of the horizontal axis of the image coordinate system of the right palm image.

[0102] In the above embodiment, when the palm image is a left palm image, the second vector is constructed along the positive horizontal axis of the image coordinate system of the left palm image using the first palm key point as the new starting point, thereby improving the accuracy of constructing the second vector for the left palm image. When the palm image is a right palm image, the second vector is constructed along the negative horizontal axis of the image coordinate system of the right palm image using the first palm key point as the new starting point, thereby improving the accuracy of constructing the second vector for the right palm image.

[0103] In one embodiment, the first palm key point is a point between the ring finger and the little finger of the palm in the palm image; the second palm key point is a point between the index finger and the middle finger of the palm in the palm image.

[0104] In one embodiment, as shown in FIG8 , when the palm image is a left palm image obtained by collecting the user's left hand (it can be understood that after the image is mirrored during the collection process, the left palm image is shown in FIG8 ), the terminal can use the first palm key point (i.e., key point 3 in FIG8 ) as the starting point of the first vector and the second palm key point (i.e., key point 1 in FIG8 ) as the end point of the first vector to construct the first vector. The terminal can use the direction of the first vector as the direction of the palm in the palm image to obtain the palm direction. Furthermore, the terminal can use key point 3 as a new starting point and construct a second vector along the positive direction of the horizontal axis of the image coordinate system of the left palm image. It can be understood that the direction of the second vector is the positive direction of the horizontal axis, that is, the calibration direction.

[0105] In one embodiment, when the user uses his left hand for forward palm verification, as shown in Figure 9, when the palm image is a left palm map obtained by collecting the user's left hand (it can be understood that after the mirroring process of the collection process, the left palm map is shown in part (a) of Figure 9), refer to part (b) of Figure 9. The terminal can determine the collection angle θ of the palm in the left palm map based on the difference between the palm direction (i.e., the direction of the first vector) and the calibration direction for the left palm map (i.e., the direction of the second vector).

[0106] In one embodiment, when a user uses their left hand for non-forward palm verification, as shown in FIG10 , if the palm image is a left palm image captured for the user's left hand (it will be understood that after mirroring during the capture process, the left palm image is shown in part (a) of FIG10 ), referring to part (b) of FIG10 , the terminal can determine the capture angle θ of the palm in the left palm image based on the difference between the palm direction (i.e., the direction of the first vector) and the calibration direction for the left palm image (i.e., the direction of the second vector). It will be understood that the capture angle calculated when the user uses their right hand for forward palm verification is smaller than the capture angle calculated when the user uses their right hand for non-forward palm verification.

[0107] In one embodiment, as shown in FIG11 , when the palm image is a right palm image obtained by capturing the user's right hand (it can be understood that after the image is mirrored during the capture process, the right palm image is shown in FIG11 ), the terminal can use the first palm key point (i.e., key point 3 in FIG11 ) as the starting point of the first vector and the second palm key point (i.e., key point 1 in FIG11 ) as the end point of the first vector to construct a first vector. The terminal can use the direction of the first vector as the direction of the palm in the palm image to obtain the palm direction. Furthermore, the terminal can use key point 3 as a new starting point and construct a second vector along the positive direction of the horizontal axis of the image coordinate system of the right palm image. It can be understood that the direction of the second vector is the negative direction of the horizontal axis, that is, the calibration direction.

[0108] In one embodiment, when the user uses his right hand to perform forward palm verification, as shown in Figure 12, when the palm image is a right palm image obtained by capturing the user's right hand (it can be understood that after the capture process is mirrored, the right palm image is shown in part (a) of Figure 12), refer to part (b) of Figure 12, the terminal can determine the capture angle θ of the palm in the right palm image based on the difference between the palm direction (i.e., the direction of the first vector) and the calibration direction (i.e., the direction of the second vector) for the right palm image.

[0109] In one embodiment, when a user uses their right hand for non-forward palm verification, if the palm image is a right palm image captured for the user's right hand, the terminal can determine the capture angle of the palm in the right palm image based on the difference between the palm direction (i.e., the direction of the first vector) and the calibration direction for the right palm image (i.e., the direction of the second vector). It will be appreciated that the capture angle calculated when the user uses their right hand for forward palm verification is smaller than the capture angle calculated when the user uses their right hand for non-forward palm verification.

[0110] In the above embodiment, the interdigital point between the ring finger and the little finger of the palm in the palm image is used as the first palm key point, and the interdigital point between the index finger and the middle finger of the palm in the palm image is used as the second palm key point. In this way, since the first palm key point and the second palm key point are located far apart in the palm, overlapping of key points can be avoided. Therefore, the first vector is constructed based on the first palm key point and the second palm key point, which can improve the accuracy of constructing the first vector.

[0111] In one embodiment, in response to an identity authentication trigger event, at least one palm image captured with a preset capture field of view is obtained, including: in response to the identity authentication trigger event, at least one initial image captured with a preset capture field of view is obtained; for each initial image, image features of the initial image are extracted, and palm detection is performed based on the image features of the initial image to obtain a palm area frame corresponding to the initial image; and the initial image is intercepted based on the palm area frame to obtain a palm image corresponding to the initial image.

[0112] Specifically, the terminal may obtain at least one initial image captured by the image acquisition unit. It is understandable that the initial image may or may not include a palm. For each initial image, the terminal may perform feature extraction on the initial image to obtain image features of the initial image, and perform palm detection based on the image features of the initial image to obtain a palm area frame corresponding to the initial image, wherein the detected palm area frame is used to locate the area where the palm is located in the initial image. Furthermore, the terminal may intercept the initial image based on the palm area frame to obtain a palm image corresponding to the initial image and including the palm. It is understandable that if there is a palm in the initial image, the terminal may detect a palm area frame for locating the area where the palm is located in the initial image, and intercept a palm image including the palm from the initial image based on the palm area frame. If there is no palm in the initial image, the terminal cannot detect the palm area frame from the initial image.

[0113] In one embodiment, the terminal may input each initial image into a trained palm detection model. For each initial image, the terminal may extract features of the initial image using the trained palm detection model to obtain image features of the initial image, and then perform palm detection based on the image features of the initial image to obtain a palm area frame corresponding to the initial image.

[0114] In one embodiment, as shown in FIG13 , the region frame information of the palm region frame detected from the initial image includes the position information of the palm region frame, namely, the coordinates (x, y) of the palm region frame in the initial image. The region frame information of the palm region frame also includes the width w and height h of the palm region frame. Furthermore, as shown in FIG14 , the terminal can intercept the initial image based on the palm region frame to obtain a palm image corresponding to the initial image and including the palm.

[0115] In the above embodiment, at least one initial image is acquired by an image acquisition unit; for each initial image, a palm region frame corresponding to the initial image is detected; and the initial image is cropped based on the palm region frame to obtain a palm image corresponding to the initial image. Furthermore, key point detection is performed based on the cropped palm image, thereby improving the accuracy of key point detection.

[0116] In one embodiment, the palm area frame is predicted by a trained palm detection model; the method also includes: obtaining a sample image; the sample image includes a sample palm; the sample image carries a reference palm area frame calibrated for the sample palm; performing palm detection on the sample image by the palm detection model to be trained to obtain a predicted palm area frame; determining a target loss value based on the position difference and width and height difference between the predicted palm area frame and the reference palm area frame; training the palm detection model to be trained by the target loss value to obtain a trained palm detection model.

[0117] The reference palm area frame is the palm area frame used as a reference during the training process of the palm detection model to be trained. The predicted palm area frame is the palm area frame obtained by performing palm detection on the sample image using the palm detection model to be trained during the training process. The position difference is the difference between the position of the predicted palm area frame in the sample image and the position of the reference palm area frame in the sample image. The position difference includes the difference between the horizontal coordinate of the predicted palm area frame in the sample image and the horizontal coordinate of the reference palm area frame in the sample image, and the difference between the vertical coordinate of the predicted palm area frame in the sample image and the vertical coordinate of the reference palm area frame in the sample image. The width and height difference includes the width difference and the height difference. The width difference is the difference between the width of the predicted palm area frame and the width of the reference palm area frame. The height difference is the difference between the height of the predicted palm area frame and the height of the reference palm area frame. It can be understood that the target loss value in this embodiment takes into account both the position difference and the width and height difference between the predicted palm area frame and the reference palm area frame.

[0118] Specifically, the terminal can obtain a sample image including a sample palm, and the sample image carries a reference palm area frame pre-calibrated for the sample palm. The terminal can input the sample image into the palm detection model to be trained, so as to perform palm detection on the sample image through the palm detection model to be trained, and output the predicted palm area frame corresponding to the sample image. It can be understood that the predicted palm area frame is used to locate the area where the sample palm is located in the sample image. The terminal can determine the target loss value based on the position difference and width and height difference between the predicted palm area frame and the reference palm area frame, and iteratively train the palm detection model to be trained in the direction of reducing the target loss value until the iteration stop condition is met to obtain a trained palm detection model.

[0119] In one embodiment, the iteration stopping condition may be at least one of the following: the number of iterations reaches a preset number of iterations or the target loss value is less than a preset loss value.

[0120] In one embodiment, the terminal may obtain an initial sample image including at least one sample palm, and calibrate a reference palm region frame for each sample palm in the initial sample image to obtain a sample image.

[0121] In one embodiment, the terminal may perform weighted processing on the position difference and width and height difference between the predicted palm area frame and the reference palm area frame, and directly use the weighted result as the target loss value.

[0122] In the above embodiment, palm detection is performed on the sample image using the palm detection model to be trained to obtain a predicted palm area frame, and a target loss value is determined based on the position difference and width and height difference between the predicted palm area frame and the reference palm area frame in the sample image. The palm detection model to be trained is then trained using the target loss value to obtain a trained palm detection model, which can improve the key point detection accuracy of the trained palm detection model.

[0123] In one embodiment, obtaining a sample image includes: obtaining an initial sample image and dividing the initial sample image into a plurality of image blocks; calibrating at least one reference palm region frame for each image block in the initial sample image to obtain a sample image.

[0124] Specifically, the terminal may obtain at least one initial sample image. For each initial sample image, the terminal may segment the initial sample image to obtain multiple image blocks corresponding to the initial sample image. Furthermore, the terminal may calibrate at least one reference palm region frame for each image block in the initial sample image to obtain a sample image corresponding to the initial sample image.

[0125] In the above embodiment, a sample image is obtained by dividing an initial sample image into multiple image blocks and calibrating at least one reference palm region frame for each image block in the initial sample image. Because the sample image is calibrated with multiple reference palm region frames, additional reference information can be added while retaining the number of sample images. Therefore, by using the sample images to train a palm detection model to be trained, a trained palm detection model is obtained, which can improve the key point detection accuracy of the trained palm detection model.

[0126] In one embodiment, the sample image also carries a reference confidence of a reference palm area box; a target loss value is determined based on the position difference and width and height difference between the predicted palm area box and the reference palm area box, including: determining a first loss value based on the position difference and width and height difference between the predicted palm area box and the reference palm area box; determining a predicted confidence of the predicted palm area box based on the area overlap between the predicted palm area box and the reference palm area box; the area overlap is positively correlated with the predicted confidence; determining a second loss value based on the confidence difference between the predicted confidence and the reference confidence; and determining a target loss value based on the first loss value and the second loss value.

[0127] Among them, the area overlap is the overlap between the area corresponding to the predicted palm area frame and the area corresponding to the reference palm area frame. The area overlap is positively correlated with the prediction confidence, that is, the higher the area overlap, the higher the prediction confidence, and the lower the area overlap, the lower the prediction confidence. The reference confidence is the confidence of the reference palm area frame used as a reference during the training process of the palm detection model to be trained. The prediction confidence is the confidence of the predicted palm area frame obtained by performing palm detection on the sample image through the palm detection model to be trained during the training process. The first loss value is a loss value determined based on the position difference and width and height difference between the predicted palm area frame and the reference palm area frame. The second loss value is a loss value determined based on the confidence difference between the predicted confidence and the reference confidence. It can be understood that the target loss value in this embodiment takes into account both the first loss value and the second loss value.

[0128] In one embodiment, if a sample palm exists in the predicted palm region frame, the terminal may determine the prediction confidence of the predicted palm region frame based on the degree of overlap between the predicted palm region frame and the reference palm region frame. If no sample palm exists in the predicted palm region frame, the prediction confidence of the predicted palm region frame is determined to be zero.

[0129] In one embodiment, the area overlap may include the area intersection-and-union ratio between the area corresponding to the predicted palm area frame and the area corresponding to the reference palm area frame. It can be understood that the area intersection-and-union ratio is the ratio between the area intersection and the area union. Among them, the area intersection is the intersection between the area corresponding to the predicted palm area frame and the area corresponding to the reference palm area frame. The area union is the union between the area corresponding to the predicted palm area frame and the area corresponding to the reference palm area frame. The terminal can determine the prediction confidence of the predicted palm area frame based on the area intersection-and-union ratio between the predicted palm area frame and the reference palm area frame. Among them, the area intersection-and-union ratio is positively correlated with the prediction confidence.

[0130] In one embodiment, the terminal may perform weighted processing based on the first loss value and the second loss value, and directly use the weighted result as the target loss value.

[0131] In the above embodiment, a first loss value is determined by the positional differences and width / height differences between the predicted palm region frame and the reference palm region frame. A second loss value is then determined based on the confidence difference between the predicted confidence level and the reference confidence level. A more comprehensive target loss value is then determined based on the first and second loss values. The target loss value is then used to train the palm detection model to be trained, resulting in a trained palm detection model. This can further improve the key point detection accuracy of the trained palm detection model.

[0132] In one embodiment, a target loss value is determined based on a first loss value and a second loss value, including: performing category probability prediction on an object in a predicted palm area box by a palm detection model to be trained to obtain a palm category probability that the object in the predicted palm area box belongs to a palm; determining a third loss value based on a difference between the predicted palm category probability and the palm category probability of a sample palm in a reference palm area box; and determining a target loss value based on the first loss value, the second loss value, and the third loss value.

[0133] The third loss value is determined based on the difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm region frame. It can be understood that the target loss value in this embodiment takes into account the first loss value, the second loss value, and the third loss value.

[0134] Specifically, the object in the predicted palm area frame may be a palm or may not be a palm. The terminal can use the palm detection model to be trained to perform a category probability prediction on the object in the predicted palm area frame to obtain a palm category probability that the object in the predicted palm area frame belongs to a palm. Furthermore, the terminal can determine a third loss value based on the difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm area frame, and determine a target loss value based on the first loss value, the second loss value, and the third loss value.

[0135] In one embodiment, the terminal may perform weighted processing on the first loss value, the second loss value, and the third loss value, and directly use the weighted result as the target loss value.

[0136] In one embodiment, the target loss value can be calculated using the following loss function:

[0137] Among them, L is the target loss value, s 2 is the number of image blocks in the sample image, B is the number of reference palm region boxes calibrated for each image block. ij represents the jth reference palm region box of the i-th image block. obj indicates that the sample palm exists in the reference palm region box, and noobj indicates that the sample palm does not exist in the reference palm region box. i Indicates the horizontal coordinate of the reference palm area frame in the sample image, y i Indicates the vertical coordinate of the reference palm area box in the sample image. Indicates the horizontal coordinate of the predicted palm area frame in the sample image, Indicates the vertical coordinate of the predicted palm area box in the sample image. It represents the difference between the horizontal coordinate of the predicted palm region frame in the sample image and the horizontal coordinate of the reference palm region frame in the sample image. It represents the difference between the vertical coordinate of the predicted palm region frame in the sample image and the vertical coordinate of the reference palm region frame in the sample image. represents the difference between the predicted position of the palm region frame in the sample image and the reference position of the palm region frame in the sample image. i Indicates the width of the reference palm area frame, h i Indicates the height of the reference palm area box. Indicates the width of the predicted palm area box, Indicates the height of the predicted palm area box. Indicates the width difference between the predicted palm region box and the reference palm region box. Indicates the height difference between the predicted palm region box and the reference palm region box. i represents the reference confidence of the reference palm area box, Prediction confidence of the predicted palm region box. The confidence difference between the predicted confidence and the reference confidence. c∈classes represents the category to which the object in the palm region belongs. It can be understood that the category to which the object belongs in this application includes the palm category. If c is the palm category, then p i(c) represents the palm category probability of the sample palm in the reference palm area frame, Indicates the probability of the palm category where the object in the predicted palm region box belongs to the palm. Represents the difference between the palm category probability and the palm category probability of the sample palm in the reference palm region frame. λ1, λ2, and λ3 are pre-set constants.

[0138] In the above embodiment, the third loss value is determined by the difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm region frame. A more comprehensive target loss value is then determined based on the first, second, and third loss values. The target loss value is then used to train the palm detection model to be trained, resulting in a trained palm detection model. This can further improve the key point detection accuracy of the trained palm detection model.

[0139] In one embodiment, key points are detected in each palm image to obtain the palm key points of each palm image, including: for each palm image, performing feature extraction on the palm image to obtain image features of the palm image; performing key point detection based on the image features of the palm image to obtain the palm key points corresponding to the palm image.

[0140] Specifically, for each palm image, the terminal may perform convolution processing on the palm image to extract the image features of the palm image. Furthermore, the terminal may perform key point detection based on the image features of the palm image to obtain the palm key points corresponding to the palm image.

[0141] In one embodiment, for each palm image, the terminal may input the palm image into a trained palm key point detection model to perform feature extraction on the palm image using the palm key point detection model to obtain image features of the palm image. Furthermore, the terminal may perform key point detection based on the image features of the palm image to obtain the palm key points corresponding to the palm image.

[0142] In one embodiment, the number of detected palm key points may include at least two. As shown in FIG15 , after acquiring the palm image captured from the initial image, the terminal may perform key point detection on the captured palm image to obtain palm key points of the palm in the captured palm image, namely, palm key point 1, palm key point 2, palm key point 3, and palm key point 4.

[0143] In the above embodiment, by extracting the image features of the palm image, since the image features of the palm image can be used to characterize the rich feature information of the palm in the palm image, key point detection is performed based on the image features of the palm image to obtain the palm key points corresponding to the palm image, which can improve the accuracy of key point detection.

[0144] In one embodiment, key point detection is performed based on image features of a palm image to obtain palm key points corresponding to the palm image, including: performing key point detection based on the image features of the palm image to obtain initial key points corresponding to the palm image; for each initial key point, cropping the palm image with reference to the initial key point to obtain a cropped image covering the initial key point and meeting a preset size; performing feature extraction on the cropped image to obtain image features of the cropped image; performing key point detection based on the image features of the cropped image to obtain palm key points corresponding to the cropped image.

[0145] It is understood that the initial keypoints are palm keypoints obtained through preliminary keypoint detection of the palm image. The initial keypoints may not be very accurate. To obtain more accurate palm keypoints, the palm image can be cropped based on the initial keypoints to obtain a cropped image that covers the initial keypoints and meets a preset size. Advanced keypoint detection can then be performed on the cropped image to obtain palm keypoints with higher accuracy than the initial keypoints.

[0146] Specifically, the terminal may perform initial key point detection based on the image features of the palm image to obtain the initial key points corresponding to the palm image. For each initial key point, the terminal may crop the palm image according to a preset size with reference to the initial key point, obtaining a cropped image that covers the initial key point and meets the preset size. It is understood that the size of the cropped image is less than or equal to the size of the palm image. Furthermore, the terminal may perform feature extraction on the cropped image to obtain image features of the cropped image, and perform advanced key point detection based on the image features of the cropped image to obtain the palm key points corresponding to the cropped image.

[0147] In one embodiment, for each palm image, the terminal may input the palm image into a trained palm key point detection model to perform feature extraction on the palm image through the palm key point detection model to obtain the image features of the palm image, and perform initial key point detection based on the image features of the palm image to obtain the initial key points corresponding to the palm image. For each initial key point, the terminal may refer to the initial key point and crop the palm image according to a preset size to obtain a cropped image that covers the initial key point and meets the preset size. Furthermore, the terminal may perform feature extraction on the cropped image to obtain the image features of the cropped image, and then input the palm image into the trained palm key point detection model to perform advanced key point detection based on the image features of the cropped image through the palm key point detection model to output the palm key points corresponding to the cropped image.

[0148] In the above embodiment, initial keypoint detection is performed based on the image features of the palm image to obtain the initial keypoints corresponding to the palm image. However, the initial keypoints obtained by initial detection may not be very accurate. Therefore, for each initial keypoint, the palm image is cropped with reference to the initial keypoint to obtain a cropped image that covers the initial keypoint and meets a preset size. Furthermore, by extracting the image features of the cropped image and performing advanced keypoint detection based on the image features of the cropped image to obtain the palm keypoints corresponding to the cropped image, the accuracy of keypoint detection can be further improved.

[0149] As shown in FIG16 , in one embodiment, an identity authentication method is provided. This embodiment uses the method applied to the terminal 102 in FIG1 as an example for explanation. The method specifically includes the following steps:

[0150] Step 1602 : obtaining an initial sample image and dividing the initial sample image into a plurality of image blocks; calibrating at least one reference palm region frame for each image block in the initial sample image to obtain a sample image.

[0151] Step 1604 , obtaining a sample image; the sample image includes a sample palm; the sample image carries a reference palm region frame calibrated for the sample palm, and a reference confidence of the reference palm region frame.

[0152] Step 1606: Perform palm detection on the sample image using the palm detection model to be trained to obtain a predicted palm region frame.

[0153] Step 1608 : Determine a first loss value based on the position difference and width and height difference between the predicted palm area frame and the reference palm area frame.

[0154] Step 1610: Determine the prediction confidence of the predicted palm area frame based on the area overlap between the predicted palm area frame and the reference palm area frame; and determine a second loss value based on the confidence difference between the predicted confidence and the reference confidence.

[0155] Step 1612 , performing category probability prediction on the object in the predicted palm region frame using the palm detection model to be trained, and obtaining a palm category probability that the object in the predicted palm region frame belongs to a palm.

[0156] Step 1614 : Determine a third loss value according to the difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm region frame.

[0157] Step 1616: Determine a target loss value based on the first loss value, the second loss value, and the third loss value, and train the palm detection model to be trained using the target loss value to obtain a trained palm detection model.

[0158] Step 1618: Acquire at least one initial image captured by the image capture unit.

[0159] Step 1620: For each initial image, the image features of the initial image are extracted using the trained palm detection model, and palm detection is performed based on the image features of the initial image to obtain a palm area frame corresponding to the initial image; the palm area frame is used to locate the area where the palm is located in the initial image.

[0160] Step 1622 , intercepting the initial image based on the palm area frame to obtain a palm image corresponding to the initial image, and performing feature extraction on each palm image to obtain image features of the palm image.

[0161] Step 1624, perform key point detection based on the image features of the palm image to obtain palm key points corresponding to the palm image; the palm key points include a first palm key point and a second palm key point; the first palm key point is the interdigital point between the ring finger and the little finger of the palm in the palm image; the second palm key point is the interdigital point between the index finger and the middle finger of the palm in the palm image.

[0162] Step 1626: construct a first vector based on the first palm key point and the second palm key point, and use the direction of the first vector as the direction of the palm in the palm image to obtain the palm direction; the starting point of the first vector is the first palm key point.

[0163] Step 1628: When the palm image is a left palm image, a second vector is constructed along the first horizontal direction with the first palm key point as a new starting point.

[0164] Step 1630: When the palm image is a right palm image, a second vector is constructed along the second horizontal direction with the first palm key point as a new starting point; the first horizontal direction and the second horizontal direction are opposite horizontal directions, and the direction of the second vector is used as the calibration direction of the image acquisition unit.

[0165] Step 1632: for each palm image, determine a preset calibration direction in the acquisition field of view, and determine the acquisition angle of the palm in the palm image according to the angle between the palm direction of the palm in the palm image and the calibration direction.

[0166] Step 1634 , screening palm images whose acquisition angles fall within a preset acquisition angle range from the at least one palm image to obtain a target palm image, and performing resource transfer processing based on the palm print information in the target palm image.

[0167] This application also provides an application scenario that utilizes the aforementioned identity verification method. Specifically, the identity verification method can be applied to a palm verification payment scenario in a supermarket. It is understood that after shopping in a supermarket, a customer needs to proceed to the checkout counter to check out. The customer can choose to perform palm verification payment at the palm verification device. It is understood that the palm verification device is deployed with a trained palm detection model. The palm detection model training steps may include: obtaining an initial sample image and dividing the initial sample image into multiple image blocks; calibrating at least one reference palm region frame for each image block in the initial sample image to obtain a sample image. A sample image is obtained; the sample image includes a sample palm; the sample image carries a reference palm region frame calibrated for the sample palm and a reference confidence score for the reference palm region frame. Palm detection is performed on the sample image using the trained palm detection model to obtain a predicted palm region frame. A first loss value is determined based on the position difference and width / height difference between the predicted palm region frame and the reference palm region frame. The prediction confidence of the predicted palm region frame is determined based on the degree of overlap between the predicted palm region frame and the reference palm region frame; a second loss value is determined based on the confidence difference between the predicted confidence and the reference confidence. The palm detection model to be trained performs a category probability prediction on the object in the predicted palm region frame to obtain a palm category probability that the object in the predicted palm region frame belongs to a palm. A third loss value is determined based on the difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm region frame. A target loss value is determined based on the first, second, and third loss values, and the target loss value is used to train the palm detection model to be trained to obtain a trained palm detection model.

[0168] When a customer places their palm over the palm verification device, the palm verification device captures at least one initial image captured by the image acquisition unit. It is understood that the initial image may or may not include the customer's palm. For each initial image, the trained palm detection model extracts image features from the initial image. Palm detection is performed based on the image features of the initial image to obtain a palm region frame corresponding to the initial image. The palm region frame is used to locate the area within the initial image where the palm is located. The initial image is captured based on the palm region frame to obtain a palm image corresponding to the initial image. For each palm image, feature extraction is performed on the palm image to obtain image features of the palm image. Keypoint detection is performed based on the image features of the palm image to obtain palm keypoints corresponding to the palm image. The palm keypoints include a first palm keypoint and a second palm keypoint. The first palm keypoint is located between the ring finger and pinky finger of the palm in the palm image; the second palm keypoint is located between the index finger and middle finger of the palm in the palm image.

[0169] The palm verification device constructs a first vector based on a first palm key point and a second palm key point, and uses the direction of the first vector as the direction of the palm in the palm image to obtain a palm direction; the starting point of the first vector is the first palm key point. If the palm image is of a left palm, a second vector is constructed along a first horizontal direction with the first palm key point as the new starting point. If the palm image is of a right palm, a second vector is constructed along a second horizontal direction with the first palm key point as the new starting point; the first and second horizontal directions are opposite horizontal directions, and the direction of the second vector serves as a calibration direction for the image acquisition unit. For each palm image, a preset calibration direction within the acquisition field of view is determined, and the acquisition angle of the palm in the palm image is determined based on the angle between the palm direction of the palm in the palm image and the calibration direction. Palm images whose acquisition angles fall within the preset acquisition angle range are screened from at least one palm image to obtain a target palm image of the customer, and palm verification payment processing is performed based on the palmprint information in the target palm image of the customer.

[0170] It can be understood that the present application detects the palm key points of the collected palm image, and automatically determines the palm collection angle based on the detected palm key points, and then filters out the target palm image of the customer used for palm verification payment from the collected palm images based on the collection angle, so as to avoid the situation where the cashier mistakenly swipes the customer when indicating the palm verification at the software level, thereby reducing the cost of using palm verification equipment in supermarket shopping scenarios.

[0171] The present application also provides an application scenario, which applies the above-mentioned identity authentication method. Specifically, the identity authentication method can be applied to a palm verification payment scenario where friends share the cost of a meal. It is understandable that after a number of friends have a meal together, they usually need to share the cost of the meal. The identity authentication method of the present application can realize palm verification payment between friends. In this way, the present application detects the key points of the palm of the collected palm image, and automatically determines the acquisition angle of the palm based on the detected key points of the palm, and then filters out the target palm image of the friend for palm verification payment from each collected palm image based on the acquisition angle, so as to avoid the situation where the verification initiator mistakenly pays when prompting the friend to verify the palm at the software level, thereby reducing the cost of using the palm verification device in the palm verification payment scenario between friends.

[0172] It should be understood that, although the various steps in the flow chart of the above-mentioned embodiments are shown in sequence, these steps are not necessarily performed in sequence. Unless clearly stated herein, the execution of these steps does not have strict order restrictions, and these steps can be performed in other sequences. Moreover, at least a portion of the steps in the above-mentioned embodiments may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0173] In one embodiment, as shown in FIG17 , an identity authentication device 1700 is provided, which specifically includes:

[0174] The acquisition module 1702 is configured to acquire at least one palm image captured in a preset capture field of view in response to an identity authentication triggering event.

[0175] The detection module 1704 is configured to detect key points in each palm image to obtain palm key points of each palm image.

[0176] Determination module 1706 is used to determine the palm direction of the palm in each palm image based on the palm key points of each palm image; for each palm image, determine the preset calibration direction in the acquisition field of view, and determine the acquisition angle of the palm in the palm image based on the angle between the palm direction of the palm in the palm image and the calibration direction.

[0177] The screening module 1708 is configured to screen the palm images whose acquisition angles are within a preset acquisition angle range from the at least one palm image to obtain a target palm image.

[0178] The identity authentication module 1710 is used to perform identity authentication based on the target palm image.

[0179] In one embodiment, the determination module 1706 is also used to, for each palm image, if the palm key points of the palm image include a first palm key point and a second palm key point, construct a first vector based on the first palm key point and the second palm key point, and use the direction of the first vector as the palm direction of the palm in the palm image.

[0180] In one embodiment, the determination module 1706 is also used to, for each palm image, if the palm key points of the palm image include a first palm key point and a second palm key point, use the first palm key point or the second palm key point as the starting point, construct a second vector along the horizontal direction, and use the direction of the second vector as the preset calibration direction in the acquisition field of view.

[0181] In one embodiment, the determination module 1706 is further used to construct a second vector along the first horizontal direction with the first palm key point or the second palm key point as the starting point when the palm in the palm image is identified as the left palm; and to construct a second vector along the second horizontal direction with the first palm key point or the second palm key point as the starting point when the palm in the palm image is identified as the right palm; wherein the first horizontal direction and the second horizontal direction are opposite horizontal directions.

[0182] In one embodiment, the determination module 1706 is further used to, when it is recognized that the palm in the palm image is the left palm, construct a second vector along the positive direction of the horizontal axis of the image coordinate system of the palm image with the first palm key point as the starting point; and when it is recognized that the palm in the palm image is the right palm, construct a second vector along the negative direction of the horizontal axis of the image coordinate system of the palm image with the first palm key point as the starting point.

[0183] In one embodiment, the first palm key point is a point between the ring finger and the little finger of the palm in the palm image; the second palm key point is a point between the index finger and the middle finger of the palm in the palm image.

[0184] In one embodiment, the acquisition module 1702 is also used to obtain at least one initial image captured with a preset capture field of view in response to an identity authentication trigger event; for each initial image, extract the image features of the initial image, and perform palm detection based on the image features of the initial image to obtain a palm area frame corresponding to the initial image; and intercept the initial image based on the palm area frame to obtain a palm image corresponding to the initial image.

[0185] In one embodiment, the palm region frame is predicted by a trained palm detection model. Apparatus 1700 further includes: a training module configured to obtain a sample image; the sample image includes a sample palm; the sample image carries a reference palm region frame calibrated for the sample palm; perform palm detection on the sample image using the palm detection model to be trained to obtain a predicted palm region frame; determine a target loss value based on the position difference and width / height difference between the predicted palm region frame and the reference palm region frame; and train the palm detection model to be trained using the target loss value to obtain a trained palm detection model.

[0186] In one embodiment, the training module is further configured to obtain an initial sample image and divide the initial sample image into a plurality of image blocks; and calibrate at least one reference palm region frame for each image block in the initial sample image to obtain a sample image.

[0187] In one embodiment, the sample image also carries a reference confidence of a reference palm area box; the training module is further used to determine a first loss value based on the position difference and width and height difference between the predicted palm area box and the reference palm area box; determine the prediction confidence of the predicted palm area box based on the area overlap between the predicted palm area box and the reference palm area box; the area overlap is positively correlated with the prediction confidence; determine a second loss value based on the confidence difference between the prediction confidence and the reference confidence; and determine a target loss value based on the first loss value and the second loss value.

[0188] In one embodiment, the training module is also used to predict the category probability of the object in the predicted palm area box through the palm detection model to be trained, and obtain the palm category probability that the object in the predicted palm area box belongs to the palm; determine the third loss value based on the difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm area box; determine the target loss value based on the first loss value, the second loss value and the third loss value.

[0189] In one embodiment, the detection module 1704 is further configured to perform feature extraction on each palm image to obtain image features of the palm image; and perform key point detection based on the image features of the palm image to obtain palm key points corresponding to the palm image.

[0190] In one embodiment, the detection module 1704 is also used to perform key point detection based on the image features of the palm image to obtain initial key points corresponding to the palm image; for each initial key point, the palm image is cropped with reference to the initial key point to obtain a cropped image covering the initial key point and meeting a preset size; feature extraction is performed on the cropped image to obtain image features of the cropped image; key point detection is performed based on the image features of the cropped image to obtain palm key points corresponding to the cropped image.

[0191] In one embodiment, the identity authentication module 1710 is further configured to perform a payment operation in response to successful identity authentication based on the biometric information in the target palm image.

[0192] The aforementioned identity verification device, in response to an identity verification trigger event, obtains at least one palm image captured within a preset capture field of view, detects key points in each palm image, and obtains palm key points for each palm image. Based on the palm key points, the device determines the palm orientation of each palm image. For each palm image, the device also determines a preset calibration direction within the capture field of view. The device then determines the palm capture angle of the palm image based on the angle between the palm orientation of the palm image and the calibration direction. This automatically selects palm images from the at least one palm image whose capture angles fall within a preset capture angle range for representing the payee, obtains a target palm image, and performs identity verification based on the target palm image. This device can ensure that authentication is only performed correctly when the palm is in a specific orientation, thus avoiding misjudgments.

[0193] Each module in the aforementioned identity verification device may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0194] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be shown in Figure 18. The computer device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via wired or wireless communication, and the wireless communication may be implemented via Wi-Fi, a mobile cellular network, NFC (near field communication), or other technologies. When executed by the processor, the computer program implements an authentication method. The display unit of the computer device is used to form a visually visible image, and can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse, etc.

[0195] Those skilled in the art will understand that the structure shown in Figure 18 is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0196] In one embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0197] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0198] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0199] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions.

[0200] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0201] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0202] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. An identity authentication method, performed by a computer device, comprising: In response to an identity authentication triggering event, acquiring at least one palm image captured with a preset capture field of view; Detect key points in each of the palm images respectively to obtain palm key points of each of the palm images; Determine the palm direction of the palm in each of the palm images according to the palm key points of each of the palm images; For each of the palm images, respectively determine a preset calibration direction in the acquisition field of view, and determine an acquisition angle of the palm in the palm image according to an angle between the palm direction of the palm in the palm image and the calibration direction; Filtering palm images whose acquisition angles are within a preset acquisition angle range from the at least one palm image to obtain a target palm image; and Authentication is performed based on the target palm image.

2. The method according to claim 1, wherein determining the palm direction of each palm image according to the palm key points of each palm image comprises: For each of the palm images, if the palm key points of the palm image include a first palm key point and a second palm key point, a first vector is constructed based on the first palm key point and the second palm key point, and the direction of the first vector is used as the palm direction of the palm in the palm image.

3. The method according to claim 2, wherein for each of the palm images, determining a preset calibration direction in the acquisition field of view, and determining the acquisition angle of the palm in the palm image according to the angle between the palm direction of the palm in the palm image and the calibration direction, comprises: For each of the palm images, if the palm key points of the palm image include a first palm key point and a second palm key point, take the first palm key point or the second palm key point as a starting point, construct a second vector along the horizontal direction, and use the direction of the second vector as a preset calibration direction in the acquisition field of view.

4. The method according to claim 3, wherein taking the first palm key point or the second palm key point as a starting point and constructing a second vector along a horizontal direction comprises: When it is recognized that the palm in the palm image is a left palm, taking the first palm key point or the second palm key point as a starting point, constructing a second vector along a first horizontal direction; When it is recognized that the palm in the palm image is the right palm, taking the first palm key point or the second palm key point as a starting point, constructing a second vector along a second horizontal direction; The first horizontal direction and the second horizontal direction are opposite horizontal directions.

5. The method according to claim 4, wherein when the palm in the palm image is identified as a left palm, taking the first palm key point or the second palm key point as a starting point, constructing a second vector along a first horizontal direction comprises: When it is recognized that the palm in the palm image is a left palm, taking the first palm key point as a starting point, constructing a second vector along the positive direction of the horizontal axis of the image coordinate system of the palm image; When the palm in the palm image is identified as the right palm, taking the first palm key point or the second palm key point as a starting point and constructing a second vector along a second horizontal direction includes: When it is recognized that the palm in the palm image is the right palm, a second vector is constructed along the negative direction of the horizontal axis of the image coordinate system of the palm image with the first palm key point as the starting point.

6. According to the method described in any one of claims 2-5, the first palm key point is the finger gap point between the ring finger and the little finger of the palm in the palm image; the second palm key point is the finger gap point between the index finger and the middle finger of the palm in the palm image.

7. The method according to any one of claims 1 to 6, wherein in response to an identity authentication triggering event, acquiring at least one palm image captured in a preset capture field of view comprises: In response to an identity authentication triggering event, acquiring at least one initial image acquired with a preset acquisition field of view; For each initial image, extract image features of the initial image, and perform palm detection based on the image features of the initial image to obtain a palm area frame corresponding to the initial image; The initial image is intercepted based on the palm area frame to obtain a palm image corresponding to the initial image.

8. The method according to claim 7, wherein the palm region frame is predicted by a trained palm detection model; the method further comprising: Acquire a sample image; the sample image includes a sample palm; The sample image carries a reference palm area frame calibrated for the sample palm; Performing palm detection on the sample image using a palm detection model to be trained to obtain a predicted palm area frame; Determining a target loss value according to a position difference and a width and height difference between the predicted palm area frame and the reference palm area frame; The palm detection model to be trained is trained using the target loss value to obtain the trained palm detection model.

9. The method according to claim 8, wherein acquiring the sample image comprises: Acquire an initial sample image, and divide the initial sample image into a plurality of image blocks; At least one reference palm area frame is calibrated for each image block in the initial sample image to obtain a sample image.

10. The method according to claim 8, wherein the sample image also carries a reference confidence of the reference palm area frame; The determining a target loss value according to a position difference and a width and height difference between the predicted palm area frame and the reference palm area frame includes: Determining a first loss value according to a position difference and a width and height difference between the predicted palm area frame and the reference palm area frame; Determining a prediction confidence of the predicted palm region frame according to a region overlap between the predicted palm region frame and the reference palm region frame; the region overlap is positively correlated with the prediction confidence; Determining a second loss value according to a confidence difference between the prediction confidence and the reference confidence; A target loss value is determined according to the first loss value and the second loss value.

11. The method according to claim 10, wherein determining a target loss value according to the first loss value and the second loss value comprises: Performing category probability prediction on the object in the predicted palm area frame by using the palm detection model to be trained to obtain a palm category probability that the object in the predicted palm area frame belongs to the palm; Determine a third loss value according to a difference between the predicted palm category probability and the palm category probability of the sample palm in the reference palm region frame; A target loss value is determined according to the first loss value, the second loss value and the third loss value.

12. The method according to any one of claims 1 to 11, wherein detecting key points in each of the palm images to obtain the palm key points of each of the palm images comprises: For each palm image, extract features of the palm image to obtain image features of the palm image; Key point detection is performed according to the image features of the palm image to obtain palm key points corresponding to the palm image.

13. The method according to claim 12, wherein the step of performing key point detection according to the image features of the palm image to obtain the palm key points corresponding to the palm image comprises: Perform key point detection according to the image features of the palm image to obtain initial key points corresponding to the palm image; For each initial key point, the palm image is cropped with reference to the initial key point to obtain a cropped image covering the initial key point and meeting a preset size; Performing feature extraction on the cropped image to obtain image features of the cropped image; Key point detection is performed according to the image features of the cropped image to obtain palm key points corresponding to the cropped image.

14. The method according to any one of claims 1 to 13, further comprising: In response to successful identity authentication based on the biometric information in the target palm image, a payment operation is performed.

15. An identity verification device, comprising: An acquisition module, configured to acquire at least one palm image acquired with a preset acquisition field of view in response to an identity authentication triggering event; A detection module, used for detecting key points in each of the palm images respectively, to obtain palm key points of each of the palm images; A determination module, configured to determine the palm direction of the palm in each of the palm images according to the palm key points of each of the palm images; for each of the palm images, determine the preset calibration direction in the acquisition field of view, and determine the acquisition angle of the palm in the palm image according to the angle between the palm direction of the palm in the palm image and the calibration direction; A screening module, configured to screen palm images whose acquisition angles are within a preset acquisition angle range from the at least one palm image, to obtain a target palm image; and The identity verification module is used to verify the identity of the target according to the target palm image.

16. A computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 14 when executing the computer program.

17. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 14.

18. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 14 are implemented.

Images