Iris recognition method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of image processing, in particular to an iris recognition method and device, electronic equipment and a storage medium, which are used for improving the accuracy of iris recognition. The method comprises the following steps: acquiring an iris image group collected for a to-be-recognized object; the iris image group comprises a plurality of iris images collected at different collection view angles for the same eye region of the to-be-recognized object; in the iris image group, performing stereo matching between every two iris images to obtain a corresponding disparity map; based on each obtained disparity map, positioning a pupil edge from any iris image contained in the iris image group; determining an iris region in any iris image according to the pupil edge, and performing feature extraction on any iris region to obtain corresponding iris features; and performing identity recognition on the to-be-recognized object based on the iris features. The pupil edge is positioned through stereo matching, the edge positioning error is reduced, and the iris recognition accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an iris recognition method, device, electronic device and storage medium. Background Art

[0002] Iris recognition is a biometric technology that analyzes the texture of an individual's iris. Highly reliable and unique, iris recognition is widely used in security authentication and identity verification scenarios.

[0003] Iris recognition methods in related technologies mainly locate pupil edges through edge detection, morphological processing, and other methods; then, based on the located pupil edges, the iris area is determined for iris recognition.

[0004] However, when pupil size changes, light changes, or pupil edge is blurred, the edge positioning effect obtained by the above method is poor, which affects the accuracy of iris recognition.

[0005] Therefore, how to improve the accuracy of iris recognition is an urgent problem to be solved. Summary of the Invention

[0006] Embodiments of the present application provide an iris recognition method, apparatus, electronic device, and storage medium to improve the accuracy of iris recognition.

[0007] An iris recognition method provided in an embodiment of the present application includes:

[0008] Acquire an iris image group collected for the object to be identified; the iris image group includes: multiple iris images collected at different acquisition viewing angles for the same eye area of the object to be identified;

[0009] In the iris image group, stereo matching is performed between each two iris images to obtain corresponding disparity maps, wherein the disparity elements in each disparity map represent: the displacement of each corresponding point between the two corresponding iris images in a specified direction;

[0010] Locating a pupil edge from any one iris image included in the iris image group based on the obtained disparity maps;

[0011] Determining an iris region in any one of the iris images according to the pupil edge, and performing feature extraction on the any one of the iris regions to obtain corresponding iris features;

[0012] Based on the iris features, the identity of the object to be identified is identified.

[0013] An iris recognition device provided in an embodiment of the present application includes:

[0014] An image acquisition unit is configured to acquire an iris image group collected for an object to be identified; the iris image group comprises: a plurality of iris images collected at different acquisition viewing angles for the same eye region of the object to be identified;

[0015] a stereo matching unit, configured to perform stereo matching between each two iris images in the iris image group to obtain a corresponding disparity map, wherein the disparity elements in each disparity map represent the displacement of corresponding points between the two iris images in a specified direction;

[0016] a pupil locating unit, configured to locate a pupil edge from any one iris image included in the iris image group based on the obtained disparity maps;

[0017] an iris feature extraction unit, configured to determine an iris region in any one iris image according to the pupil edge, and perform feature extraction on the any one iris region to obtain corresponding iris features;

[0018] The identification unit is used to identify the object to be identified based on the iris feature.

[0019] Optionally, the stereo matching unit is specifically configured to:

[0020] Eliminating the parallax between the two iris images by correcting the two iris images;

[0021] Extract pupil feature points from each rectified iris image respectively;

[0022] Through stereo matching, corresponding points are determined from the extracted pupil feature points;

[0023] Based on the determined positions of the corresponding points in the corresponding iris images, a disparity map corresponding to the two iris images is obtained.

[0024] Optionally, the pupil positioning unit is specifically used to:

[0025] If a disparity map is obtained, locating a pupil edge from any one iris image included in the iris image group based on the disparity map;

[0026] If multiple disparity maps are obtained, the multiple disparity maps are image fused, and based on the fused disparity map, the pupil edge is located from any iris image included in the iris image group; alternatively, based on each disparity map, a pupil edge is located from any iris image corresponding to the disparity map, and the multiple determined pupil edges are fused to obtain a fused pupil edge.

[0027] Optionally, the pupil positioning unit is specifically used to:

[0028] Based on a disparity map, the pupil edge is located from an iris image as follows:

[0029] Extracting gradient information from the disparity map using an edge detection operator to obtain a gradient map corresponding to the disparity map; a gradient element in the gradient map represents: a rate of change of the grayscale value of each pixel in the disparity map;

[0030] Determining a pupil edge point from the disparity map based on a preset gradient threshold;

[0031] The pupil edge is located from the one iris image according to the determined pupil edge point.

[0032] Optionally, the iris feature extraction unit is specifically used to:

[0033] Setting a preset margin around the pupil edge in any one of the iris images;

[0034] An annular area determined based on the pupil edge and the preset margin is used as the iris area in any one iris image.

[0035] Optionally, the iris feature extraction unit is specifically configured to perform at least one of the following steps:

[0036] Extracting features from the iris region using filters of different scales and directions to obtain corresponding response values; combining the response values extracted by each filter to form iris features corresponding to the iris region;

[0037] By comparing the grayscale values of each pixel point in the iris area with the corresponding neighboring pixel points, the iris features corresponding to the iris area are extracted.

[0038] Optionally, the iris feature extraction unit is further configured to enhance the contrast of the iris region before performing feature extraction on the iris region to obtain corresponding iris features.

[0039] Optionally, the iris feature extraction unit is further configured to, before extracting features from the iris region to obtain corresponding iris features, process at least one disparity map in the following manner to update the iris region:

[0040] Extracting gradient information from the disparity map using different edge detection operators to obtain multiple gradient maps corresponding to the disparity map; wherein each edge detection operator corresponds to a gradient map; and gradient elements in the gradient map represent: a rate of change of the grayscale value of each pixel in the disparity map;

[0041] determining a new pupil edge point from the disparity map based on the multiple gradient maps;

[0042] The iris region in any one iris image is updated according to the new pupil edge point.

[0043] Optionally, if there are multiple iris image groups and each iris image group corresponds to an image pattern, the stereo matching unit is further configured to:

[0044] Dividing the plurality of iris image groups into candidate iris image sets corresponding to different acquisition perspectives, wherein the candidate iris images in each candidate iris image set are iris images acquired in different image modes and at the same acquisition perspective;

[0045] For each pixel point, the corresponding pixel points on each candidate iris image in the same candidate iris image set are subjected to feature fusion to obtain a fused iris image;

[0046] The stereo matching unit is specifically used for:

[0047] Stereo matching is performed between each two fused iris images to obtain the corresponding disparity map.

[0048] Optionally, the image mode includes some or all of the following:

[0049] RGB image mode, infrared image mode, visible light image mode.

[0050] Optionally, the stereo matching unit is specifically configured to:

[0051] The grayscale values of the corresponding pixels on each candidate iris image in the same candidate iris image set are weighted averaged.

[0052] Optionally, before performing stereo matching between each two iris images to obtain corresponding disparity maps, the stereo matching unit is further configured to:

[0053] For each iris image in the iris image group, image enhancement processing is performed on the iris image in at least one manner.

[0054] An electronic device provided in an embodiment of the present application includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of any one of the above-mentioned iris recognition methods.

[0055] An embodiment of the present application provides a computer-readable storage medium, which includes a computer program. When the computer program is run on an electronic device, the computer program is used to enable the electronic device to perform the steps of any one of the above-mentioned iris recognition methods.

[0056] An embodiment of the present application provides a computer program product, which includes a computer program stored in a computer-readable storage medium; when a processor of an electronic device reads the computer program from the computer-readable storage medium, the processor executes the computer program, causing the electronic device to perform the steps of any one of the above-mentioned iris recognition methods.

[0057] The beneficial effects of this application are as follows:

[0058] The embodiments of the present application provide an iris recognition method, apparatus, electronic device, and storage medium. In the embodiments of the present application, during iris recognition, a set of iris images is collected, and stereo matching is performed between each pair of iris images in the set to ultimately locate the pupil edge. Because the present application locates the pupil edge using stereo matching technology, which can accurately locate the pupil edge under different lighting conditions, facial postures, and expressions, and is highly robust to interference factors such as noise and occlusion, locating the pupil edge in this manner can effectively reduce pupil edge positioning errors.

[0059] In summary, stereo matching technology can more accurately locate the pupil edge under complex lighting and pupil size conditions. Accurately finding the pupil edge allows for a more precise determination of the iris range, improving iris recognition accuracy. This technology also maintains good recognition results even in complex situations such as pupil size fluctuations, lighting changes, and blurred pupil edges.

[0060] In addition, stereo matching technology has high real-time performance and can quickly detect and locate pupil edges, meeting the needs of real-time applications. It can also be optimized and improved according to different application scenarios and needs, and has high scalability and flexibility.

[0061] Other features and advantages of the present application will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present application. The purposes and other advantages of the present application can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0063] Figure 1 This is an optional schematic diagram of an application scenario in an embodiment of the present application;

[0064] Figure 2 A schematic diagram of an iris image of a human eye according to an embodiment of the present application;

[0065] Figure 3 An implementation flow chart of an iris recognition method provided in an embodiment of the present application;

[0066] Figure 4 A logical diagram of stereo matching of four iris images in an embodiment of the present application;

[0067] Figure 5 is a schematic diagram of a grayscale histogram in an embodiment of the present application;

[0068] Figure 6 is a schematic diagram of two iris images in an embodiment of the present application;

[0069] Figure 7 A schematic diagram of a correction principle in an embodiment of the present application;

[0070] Figure 8 is a schematic diagram of an iris image before and after correction in an embodiment of the present application;

[0071] Figure 9 is a schematic diagram of disparity map fusion in an embodiment of the present application;

[0072] Figure 10 This is a flow chart of a pupil edge positioning method according to an embodiment of the present application;

[0073] Figure 11 A schematic diagram of a method for determining an iris area in an embodiment of the present application;

[0074] Figure 12 This is a flowchart of an iris area updating method according to an embodiment of the present application;

[0075] Figure 13 This is a schematic diagram of a process for identifying an object through feature matching in an embodiment of the present application;

[0076] Figure 14 A schematic diagram of the division logic of a candidate iris image set in an embodiment of the present application;

[0077] Figure 15This is a modular block diagram of an iris recognition process in an embodiment of the present application;

[0078] Figure 16 This is a schematic diagram of the structure of an iris recognition device in an embodiment of the present application;

[0079] Figure 17 A schematic diagram of the hardware structure of an electronic device to which an embodiment of the present application is applied;

[0080] Figure 18 A schematic diagram of the hardware structure of another electronic device to which an embodiment of the present application is applied. DETAILED DESCRIPTION

[0081] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of the technical solutions of this application, but not all of them. Based on the embodiments described in this application document, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the technical solutions of this application.

[0082] The following is an introduction to some concepts involved in the embodiments of this application.

[0083] The pupil is the small circular hole in the center of the iris in an animal or human eye. It is the passageway for light to enter the eye. Contraction of the pupillary sphincter muscle on the iris causes the pupil to constrict, while contraction of the pupillary dilator muscle causes the pupil to dilate. The dilation and constriction of the pupil control the amount of light entering the pupil.

[0084] The iris is a thin, pigmented ring-shaped membrane at the front of the eyeball, located outside the pupil. The center of the circle containing the iris is usually the center of the pupil.

[0085] Iris Recognition: A biometric technology that identifies individuals by analyzing the texture features of the iris.

[0086] Stereo Matching: A computer vision technique that compares the similarities between two stereo views to determine the positions of corresponding points in the image, thereby achieving 3D reconstruction or object positioning.

[0087] Disparity Map: A disparity map is an image that represents the horizontal displacement between corresponding points in the left and right views. Disparity maps can be used to calculate depth information for objects, enabling 3D reconstruction or object localization. In the stereo matching process, a disparity map is a representation of the matching results. Each pixel in the disparity map represents the horizontal displacement between corresponding points in the left and right views. Larger values indicate that the object is closer to the observer (such as the binocular camera described below).

[0088] Artificial Intelligence (AI) is the theory, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to perceive, reason, and make decisions. AI technology is an interdisciplinary discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Basic AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning. The technical solutions provided in the embodiments of this application primarily address computer vision and machine learning / deep learning within AI.

[0089] Computer vision (CV) is the science of making machines "see." Specifically, it refers to machine vision, where cameras and computers replace the human eye in identifying and measuring objects, performing further image processing to create images more suitable for human observation or transmission to instrumentation. As a scientific discipline, computer vision studies related theories and technologies, aiming to build artificial intelligence systems capable of extracting information from images or multidimensional data. Computer vision technologies typically include image processing, image recognition, image semantic understanding, image retrieval, video processing, video semantic understanding, video content / behavior recognition, three-dimensional object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping, and common biometric recognition technologies such as facial recognition and fingerprint recognition.

[0090] The technical solution provided by the embodiments of this application mainly relates to image recognition in computer vision technology. Specifically, the iris recognition method in this application realizes the identity recognition of the object to be recognized by recognizing the iris image collected based on the object to be recognized.

[0091] The following is a brief introduction to the design concept of the embodiment of this application:

[0092] Biometrics is a recently emerging technology that uses physiological characteristics to verify identity. Compared to traditional identification technologies (such as door keys and passwords), biometrics offers greater stability, security, and portability.

[0093] Iris recognition is a biometric technology that uses the texture characteristics of the iris to authenticate identity. Compared to other biometric technologies such as facial recognition, palm print recognition, and fingerprint recognition, iris recognition technology has unique advantages. First, the iris is highly biologically active and coexists with human life phenomena, so it is impossible to replace the iris image of a living person with a photo or video. Second, the iris is extremely stable. It is formed before birth, takes shape 6-18 months after birth, and remains unchanged throughout life. Finally, the iris is unique. The information contained in each iris is extremely random, and the iris texture of the left and right eyes of the same person will not be identical.

[0094] Specifically, iris recognition analyzes the iris texture features in an individual's eyes. Iris recognition is highly reliable and unique, and is widely used in security authentication and identity verification scenarios.

[0095] The iris recognition method in the related art collects iris images through iris acquisition equipment, and then detects the pupil edge in the iris image. Currently, the pupil edge is mainly located through edge detection, morphological processing and other methods; then, based on the located pupil edge, the iris area is determined for iris recognition.

[0096] However, when pupil size changes, light changes, or pupil edge is blurred, the edge positioning effect obtained by the above method is poor, which affects the accuracy of iris recognition.

[0097] In view of this, embodiments of the present application provide an iris recognition method, apparatus, electronic device, and storage medium. In embodiments of the present application, during iris recognition, a group of iris images is collected, and stereo matching is performed between each pair of iris images in the group to ultimately locate the pupil edge. Because the present application locates the pupil edge using stereo matching technology, which can accurately locate the pupil edge under different lighting conditions, facial postures, and expressions, and is highly robust to interference factors such as noise and occlusion, locating the pupil edge in this manner can effectively reduce pupil edge positioning errors.

[0098] In summary, stereo matching technology can more accurately locate the pupil edge under complex lighting and pupil size conditions. Accurately finding the pupil edge allows for a more precise determination of the iris range, improving iris recognition accuracy. This technology also maintains good recognition results even in complex situations such as pupil size fluctuations, lighting changes, and blurred pupil edges.

[0099] In addition, stereo matching technology has high real-time performance and can quickly detect and locate pupil edges, meeting the needs of real-time applications. It can also be optimized and improved according to different application scenarios and needs, and has high scalability and flexibility.

[0100] The preferred embodiments of the present application are described below in conjunction with the drawings in the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application and are not used to limit the present application. In addition, the embodiments and features in the embodiments of the present application can be combined with each other if there is no conflict.

[0101] The solution provided in the embodiment of the present application can be applied to iris recognition, such as iris recognition for identity recognition. The solution can be applied as a basic technology in various scenarios, including but not limited to cloud technology, artificial intelligence, smart transportation, assisted driving and other scenarios. Figure 1 , which is a schematic diagram of an application scenario in an embodiment of the present application. The application scenario diagram includes an iris acquisition device 101 and an iris recognition device 102.

[0102] The iris acquisition device 101 serves as a front-end device and can be used to acquire iris images of the eye area of the object to be identified. It may include but is not limited to contact iris acquisition devices, intelligent iris and face all-in-one machines, portable iris recognition devices, long-distance contactless iris acquisition and recognition devices, and other devices specifically used for iris image acquisition. It can also be a mobile phone, tablet computer (PAD), laptop computer, desktop computer, smart car-mounted equipment, smart voice interaction equipment, smart home appliances, smart wearable devices, aircraft, and other terminal devices with iris image acquisition functions. This article does not make specific restrictions.

[0103] Iris recognition technology is widely used in various fields due to its high security and accuracy. As technology advances, iris capture devices are also constantly improving, becoming more efficient, convenient, and adaptable to a variety of application scenarios. When selecting an iris capture device, consider its capture efficiency, accuracy, portability, and whether it meets the needs of the specific application.

[0104] like Figure 1 In the example, a binocular camera such as an iris acquisition device is used. As shown in 1011, the binocular camera is used to take a picture of the object to be identified 1012 to acquire an iris image. Figure 1 The object to be identified 1012 is a person. Of course, in addition to this, the iris recognition method in the embodiment of the present application is also applicable to other individuals with irises, such as other primates, other mammals, birds, etc., which will not be described in detail here.

[0105] The following is a simple explanation using people as an example:

[0106] like Figure 2 As shown in FIG, it is a schematic diagram of an iris image of a human eye in an embodiment of the present application. The structure of the human eye is composed of sclera, iris, pupil, lens, retina and other parts. Figure 2 As shown, the iris ( Figure 2 The gray area (in the image) is the circular area between the black pupil and the white sclera. It contains numerous interwoven details, including spots, filaments, coronae, striae, and crypts. When illuminated by infrared light of a certain wavelength (generally between 700 and 900 nanometers), the iris exhibits a radial structure extending from the inside out. These subtle features are known as iris texture. This unique characteristic has important applications in various fields.

[0107] It should be noted that the iris images in this article are just simple examples. The texture and other detailed features of the actual iris may not be reflected in the drawings in this article, but this does not mean that it does not contain these features. This is hereby explained.

[0108] The iris recognition device 102 can be any electronic device with iris recognition capabilities, such as a security access control system, an identity authentication device, etc., or a terminal device, a server, etc. Taking the iris recognition device 102 as a server as an example, it can be a standalone physical server, a server cluster or a distributed system consisting of two physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0109] If the iris collection device is a terminal device, an iris recognition-related client may also be installed on the terminal device. The client may be software (such as payment software, iris recognition software, etc.), or a web page, mini program, etc. Correspondingly, the server is a background server corresponding to the software or web page, mini program, etc., or a server specifically used for iris recognition. This application does not make specific limitations.

[0110] In actual application, the iris acquisition device 101 can send the collected iris image group containing multiple iris images to the iris recognition device 102. The iris recognition device 102 then uses the iris recognition method of the embodiment of the present application to perform stereo matching between every two iris images in the iris image group to obtain a corresponding disparity map; then, based on the obtained disparity maps, the pupil edge is located from any iris image included in the iris image group; based on the pupil edge, the iris area in any iris image is determined, and features are extracted from any iris area to obtain corresponding iris features; finally, based on the extracted iris features, the identity of the object to be identified is identified.

[0111] It should be noted that, in actual applications, when the processing capability of the iris acquisition device is sufficient, the iris recognition device 102 and the iris acquisition device 101 can be implemented by the same device, and the embodiments of the present application do not impose any restrictions on this.

[0112] In the embodiment of the present application, the iris acquisition device 101 and the iris recognition device 102 can be directly or indirectly connected to each other through one or more networks. The network can be a wired network or a wireless network. For example, the wireless network can be a mobile cellular network or a Wireless Fidelity (WIFI) network. Of course, other possible networks are also possible, and the embodiment of the present application does not limit this.

[0113] It should be noted that Figure 1 The examples shown are just for illustration. In fact, the number of iris acquisition devices and iris recognition devices is not limited and is not specifically limited in the embodiments of this application.

[0114] In an embodiment of the present application, when the number of servers is two, the two servers can form a blockchain, and the servers are nodes on the blockchain; as disclosed in the embodiment of the present application, the iris recognition method, in which the iris recognition-related data involved can be stored on the blockchain, for example, iris images, disparity maps, pupil edges, iris areas, iris features, identity recognition results, etc.

[0115] The following are some common iris recognition application scenarios:

[0116] (1) Security access control system: Iris recognition can be used in access control systems in enterprises, government agencies, laboratories and other places to ensure that only authorized personnel can enter specific areas.

[0117] (2) Banks and financial institutions: Iris recognition can be used in financial scenarios such as banks and ATMs, providing a safe and convenient way to authenticate customer identities.

[0118] (3) Unlocking electronic devices: Iris recognition can be used to unlock electronic devices such as smartphones and tablets, providing a safer and more convenient unlocking method.

[0119] (4) Attendance system: Iris recognition can be used in the attendance system of enterprises and schools to ensure that the attendance records of employees or students are accurate.

[0120] (5) Medical industry: Iris recognition can be used in medical institutions to authenticate patients and ensure the accuracy and security of medical information.

[0121] (6) Driving license test: Iris recognition can be used in driving license tests to ensure the authenticity of the candidate's identity and prevent cheating.

[0122] (7) Voting system: Iris recognition can be used in election voting systems to ensure the authenticity of voters’ identities and guarantee the fairness of elections.

[0123] (8) Smart Home: Iris recognition can be used in smart home systems to identify family members and provide personalized home services.

[0124] It should be noted that the several iris recognition scenarios listed above are just simple examples. Other iris recognition scenarios are also applicable to the embodiments of the present application and will not be described in detail here.

[0125] In addition, it can be understood that in the specific implementation of this application, related data such as iris images are involved. When the above embodiments of this application are applied to specific products or technologies, it is necessary to obtain the permission or consent of the object, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0126] The following describes the iris recognition method provided by the exemplary embodiment of the present application in combination with the application scenarios described above and with reference to the accompanying drawings. It should be noted that the above application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the implementation of the present application is not limited in this respect.

[0127] See Figure 3 FIG. 1 is a flowchart of an iris recognition method according to an embodiment of the present application. The specific implementation process of the method is as follows S31 to S35:

[0128] S31: Acquire a group of iris images collected for the object to be identified.

[0129] The iris image group includes: multiple iris images captured at different capture angles for the same eye region of the object to be identified.

[0130] Specifically, this step is mainly to collect the iris image of the object (individual) to be identified. In this application, a dedicated iris acquisition device, such as an iris recognition camera, a contact iris acquisition device, a contactless iris acquisition device, etc., can be used to capture a clear iris image.

[0131] Among them, the iris images collected in this application need to be stereo matched. Stereo matching technology determines the positions of corresponding points in the image by comparing the similarities in two or more stereo views, thereby achieving three-dimensional reconstruction or object positioning.

[0132] Therefore, it is necessary to capture iris images at different capture angles for the same eye region of the object to be identified, so as to obtain multiple iris images, wherein each capture angle corresponds to at least one iris image.

[0133] For example, using the left and right cameras of a binocular camera, left and right view images of the same scene can be captured, where the left and right cameras correspond to different acquisition angles. Therefore, by using the binocular camera to capture iris images of the same eye area of the object to be identified, iris images containing a certain overlapping area under two different acquisition angles can be obtained for matching.

[0134] For another example, a monocular camera or other camera may be used to switch different acquisition viewing angles to acquire iris images of the same eye area of the object to be identified, and so on.

[0135] It should be noted that any method of using any iris acquisition device to acquire an iris image group for an object to be identified is applicable to the embodiments of the present application and will not be described in detail here.

[0136] In order to further improve the image quality and thus improve the accuracy of subsequent pupil edge positioning, some preprocessing operations to enhance the image quality may be performed on the collected iris image.

[0137] In an optional implementation, after step S31 and before step S32, the collected iris image may be preprocessed in at least one of the following ways:

[0138] Method 1: De-noise the iris image.

[0139] Specifically, image denoising aims to remove noise from images to improve image quality and visual effects. Noise may come from factors such as sensor errors and uneven lighting during image acquisition.

[0140] In the embodiments of the present application, there are many methods for denoising an iris image, including but not limited to some or all of the following:

[0141] (1) Denoising through filters.

[0142] Such as: mean filter, Gaussian filter, median filter, bilateral filter, etc.

[0143] Among them, the mean filter smoothes the iris image and reduces noise by replacing the pixel value with the average value of the pixel's neighborhood. The Gaussian filter uses a Gaussian function as a weight to perform a weighted average on the iris image to achieve a smoothing effect. The median filter takes the current pixel as the center, takes a fixed-size neighborhood (such as 3x3, 5x5, etc.), sorts the pixel values in the neighborhood, and then takes the median as the new value of the current pixel. This method can effectively preserve image edge information while removing noise. The bilateral filter combines spatial proximity and pixel value similarity to retain edge information while removing noise.

[0144] In iris recognition, the commonly used denoising method is median filtering, which can effectively eliminate salt-and-pepper noise.

[0145] (2) Non-local mean denoising: Utilizing the self-similarity of iris images, the entire iris image is averaged to effectively preserve texture and details.

[0146] (3) Wavelet transform: Through multi-scale analysis, the iris image is decomposed into wavelets at different levels, and then the detail coefficients are thresholded, and finally the denoised iris image is reconstructed.

[0147] (IV) Anisotropic diffusion: simulates the heat conduction process and gradually smoothes the iris image through iterative calculation while preserving the edges.

[0148] (5) Total variation denoising: Remove noise by minimizing the total variation of the iris image and maintain the edge and texture information of the iris image.

[0149] (6) Low-rank matrix recovery: The iris image is regarded as a matrix and noise is removed by solving the low-rank matrix recovery problem.

[0150] (7) Deep learning methods: Utilize deep neural networks, such as Convolutional Neural Networks (CNN) and Generative Adversarial Networks (GAN), to learn prior knowledge of iris images and achieve efficient denoising.

[0151] It should be noted that the image denoising methods listed above are only examples. Other methods are also applicable to the embodiments of this application and will not be described in detail here. In addition, the selection of an appropriate denoising method depends on the noise type, image content, and application scenario. In actual applications, it may be necessary to combine multiple methods or adjust parameters to achieve the best denoising effect, which is not specifically limited in this article.

[0152] Method 2: Perform histogram equalization on the iris image.

[0153] Histogram equalization is an image enhancement method that can increase image contrast and make iris texture more visible. The basic idea is to stretch the grayscale levels that appear frequently in the image and reduce the grayscale levels that appear less frequently. This makes the image's grayscale histogram more uniform, thereby improving the overall contrast and visual quality of the image.

[0154] Specifically, the steps of performing histogram equalization on the iris image are summarized as follows:

[0155] (1) The original iris image is converted into a grayscale image, its grayscale level is determined, and a grayscale histogram is obtained statistically. The grayscale histogram represents the number distribution of pixels of each grayscale level in the iris image.

[0156] like Figure 5 As shown, it is a schematic diagram of a grayscale histogram in an embodiment of the present application. Figure 5 The horizontal axis k represents the grayscale level, and the vertical axis h(k) represents the number of pixels. Figure 5 A simple example of 4 gray levels and the number of pixels at each gray level is shown below: Figure 5 In the image, there are 4 pixels of gray level 0, 5 pixels of gray level 1, 3 pixels of gray level 2, and 4 pixels of gray level 3.

[0157] It should be noted that in reality, the grayscale level in the image will be more (usually 256), and the number of pixels in the image will also be more. Figure 5 These are just simple examples and I will not elaborate on them one by one here.

[0158] (2) Calculate the cumulative distribution function (CDF) of each gray level.

[0159] The cumulative distribution function, abbreviated as the distribution function, is the integral of the probability density function and can completely describe the probability distribution of a real random variable.

[0160] For example, there are a total of L (usually 256) gray levels in the original iris image, and the number of pixels for each gray level is n i , 0 < i < L. Then the occurrence probability of the pixel with gray level i in this image: p x (i) = p(x = i) = n i / n. Where n is the total number of pixels in this image.

[0161] The cumulative distribution function of p x is the cumulative normalized histogram of the image:

[0162] (3) Use the cumulative distribution function to map the gray values of the original iris image to new gray values, making the gray distribution of the new image more uniform.

[0163] The specific mapping formula is as follows:

[0164]

[0165] Among them, cdfmin is the minimum value of the cumulative distribution function, M and N respectively represent the number of pixels in the length and width, v is the original gray value, and h(v) is the new gray value after mapping.

[0166] In the embodiments of this application, histogram equalization can simply and effectively improve the contrast of the image. Especially for images with a small dynamic range, after histogram equalization processing, the contrast of the iris image is improved, making the details such as iris texture in the iris image clearer and making the boundary between the pupil and the iris clearer, which is beneficial to subsequent operations such as pupil edge localization and iris feature extraction.

[0167] Method 3: Gray scale stretching.

[0168] By linearly stretching the pixel values of the iris image, the contrast range is expanded, making the details of the iris image more obvious.

[0169] In addition to the methods of enhancing the image quality listed above, image enhancement can also be performed through other methods to improve the image quality. For example, sharpening, color correction, etc. can also be used to enhance the image quality.

[0170] It should be noted that each method has its own applicable scenarios and advantages, and the selection of the appropriate method depends on the specific application requirements and image characteristics. In the embodiments of the present application, the purpose of image enhancement is to improve the visual effect of the image and make it more suitable for further analysis and processing, such as subsequent pupil edge positioning and iris feature extraction. In actual applications, it may be necessary to combine multiple methods to achieve the optimal enhancement effect, which is not specifically limited here.

[0171] In the above embodiment, by improving the image quality through certain preprocessing operations, the detailed features related to the iris, pupil, etc. in the image can be made clearer, thereby effectively improving the accuracy of subsequent pupil edge positioning.

[0172] S32: In the iris image group, stereo matching is performed between every two iris images to obtain corresponding disparity maps.

[0173] The disparity elements in each disparity map represent the displacement of corresponding points between the two iris images in a specified direction.

[0174] Specifically, the designated direction is related to the placement position of the iris acquisition device.

[0175] Taking a binocular camera as an example, the disparity map is actually the difference in pixel positions when imaging the same scene from two cameras. Because the two binocular cameras are placed horizontally, the positional deviation is reflected in the horizontal direction. Of course, if the cameras are not placed horizontally, the specified direction can also be other directions. Specific situations require specific analysis and are not discussed here.

[0176] In the embodiment of the present application, when the number of iris images in the iris image group is different, the number of obtained disparity maps is also different:

[0177] If the iris image group includes only two iris images, stereo matching is performed on the two iris images to obtain a disparity map, which reflects the displacement of corresponding points in the two iris images in a specified direction (such as the horizontal direction).

[0178] If the iris image group contains three or more iris images, these iris images all correspond to the same eye area of the object to be identified, and therefore there is a certain overlap between each of them (if there is no overlap between a certain iris image and other iris images, or the overlap is not related to the eye, then this iris image can be ignored). Therefore, stereo matching can be performed between every two iris images in the iris image group to obtain multiple disparity maps, where every two iris images correspond to one disparity map.

[0179] like Figure 4 As shown, it is a logical schematic diagram of stereo matching of four iris images in an embodiment of the present application.

[0180] Assume that there are four iris images in the iris image group, which are iris image 1, iris image 2, iris image 3 and iris image 4. Stereo matching is performed between iris image 1 and iris image 2 to obtain the disparity Figure 1 , stereo matching is performed between iris image 1 and iris image 3, and the disparity can be obtained Figure 2 , stereo matching is performed between iris image 1 and iris image 4, and the disparity can be obtained Figure 3 , stereo matching is performed between iris image 2 and iris image 3, and the disparity can be obtained Figure 4 , stereo matching is performed between iris image 2 and iris image 4, and the disparity can be obtained Figure 5 , stereo matching is performed between iris image 3 and iris image 4, and the disparity can be obtained Figure 6 .

[0181] It should be noted that the disparity map in the embodiment of the present application is consistent in size with the iris image, that is, the disparity map corresponds one-to-one to the pixels in the iris image.

[0182] The following describes the process of performing stereo matching between two iris images to obtain a disparity map:

[0183] Stereo matching is a computer vision technique used to find corresponding points in two stereo images, enabling 3D reconstruction or object localization. In iris recognition, stereo matching can be used to locate the pupil edge.

[0184] An optional implementation is to implement the stereo matching process in S32 according to the following steps, including the following steps S321 to S324 ( Figure 3 not shown):

[0185] S321: Eliminate the parallax between the two iris images by correcting the two iris images.

[0186] Specifically, stereo matching requires the use of two iris images. Taking the images taken by two cameras of a binocular camera as an example, the image taken by the left camera is recorded as the left view, and the image taken by the right camera is recorded as the right view.

[0187] like Figure 6 As shown, it is a schematic diagram of two iris images in an embodiment of the present application. Figure 6 The left view in the figure is an image of one eye of the object to be identified, captured by the left camera of the binocular camera, and the right view is an image of the eye captured by the right camera of the binocular camera. Both images can be called iris images, and the two images have a certain overlapping area for matching.

[0188] Furthermore, before stereo matching, the images can be corrected to eliminate parallax between the left and right views. Parallax is caused by the positional difference between the left and right eyes, which can affect the accuracy of stereo matching. There are many correction methods, such as binocular stereo rectification (stereo rectification), perspective transformation, affine transformation, etc. Therefore, when forming the disparity map, each point at each position only represents the horizontal displacement of pixels between the left and right views.

[0189] Binocular stereo correction is an image preprocessing method used to eliminate parallax between binocular image pairs, aligning the left and right views to the same horizontal line. This simplifies the stereo matching process and improves matching accuracy. Binocular stereo correction is typically achieved by calculating the camera's intrinsic and extrinsic parameters and performing a geometric transformation on the images.

[0190] Among them, stereo correction is to strictly correspond the two images after eliminating distortion, and use the epipolar constraint to make the epipolar lines of the two images exactly on the same horizontal line. In this way, any point on one image and its corresponding point on the other image must have the same row number, and only one-dimensional search is required in the row to match the corresponding point.

[0191] Binocular correction is based on the monocular intrinsic parameter data (focal length, imaging origin, distortion coefficient) and the binocular relative position relationship (rotation matrix and translation vector) obtained after camera calibration. It eliminates distortion and aligns the left and right views respectively, making the imaging origin coordinates of the left and right views consistent, the optical axes of the two cameras parallel, the left and right imaging planes coplanar, and the epipolar lines aligned.

[0192] like Figure 7 As shown, it is a schematic diagram of a correction principle in an embodiment of the present application.

[0193] Assume that there is a point P in space, whose coordinates in the world coordinate system are Pw, its coordinates in the left camera coordinate system in the binocular camera can be expressed as: Pl, and its coordinates in the right camera coordinate system in the binocular camera can be expressed as: Pr.

[0194] like Figure 7 Obviously, the two images after correction have achieved plane coplanarity and epipolar line alignment, so that any point on one image and its corresponding point on the other image must have the same row number. In this way, only one-dimensional search in the row is needed to match the corresponding point, which improves the efficiency of subsequent corresponding point matching.

[0195] In iris recognition, the above binocular stereo correction method is used to perform stereo correction on the two iris images. Figure 8 As shown, it is a schematic diagram of an iris image before and after correction in an embodiment of the present application. Figure 8The two iris images on the left represent the left and right view images of the same eye area of the object to be identified captured by the binocular camera before correction. The two iris images after binocular stereo correction are as follows: Figure 8 Shown on the right side.

[0196] In the above embodiment, through image correction, the search range for matching corresponding points in two iris images is reduced from two dimensions to one dimension, which greatly improves the efficiency.

[0197] S322: Extract pupil feature points from each corrected iris image.

[0198] In the embodiment of the present application, it is necessary to extract pupil feature points from the corrected image for matching. Specifically, the extracted pupil feature points include but are not limited to some or all of the following:

[0199] (1) Corner point.

[0200] Corner points are a type of local feature in iris images, usually referring to points with significant angle changes. Corner points have good stability and distinguishability.

[0201] In the embodiment of the present application, algorithms for extracting corner points include Harris corner detection, Shi-Tomasi corner detection, etc., which are not specifically limited in this article.

[0202] (2) Edge points.

[0203] Among them, edge points are also a kind of local features in the image, usually referring to points in the iris image where the grayscale value changes significantly, and can be used to describe the contour and shape information of the object.

[0204] In the embodiment of the present application, the algorithms for extracting edge points include Canny edge detection, Sobel edge detection, Roberts edge detection, Prewitt edge detection, Laplacian edge detection, Laplacian of Gaussian (LoG) edge detection, etc., which are not specifically limited in this article.

[0205] S323: Determine corresponding points from the extracted pupil feature points through stereo matching.

[0206] After the feature points are extracted, a stereo matching algorithm may be used to find corresponding points in two iris images (such as the left view and the right view listed above).

[0207] The stereo matching algorithms include but are not limited to some or all of the following:

[0208] Area-based matching algorithm, feature-based matching algorithm, etc.

[0209] In the embodiment of the present application, stereo matching is mainly used for pupil positioning in iris recognition. Taking the region-based matching algorithm as an example, the positions of corresponding points are determined by comparing the similarity of pixel values in two iris images (such as the left view and right view listed above).

[0210] Specifically, pixel value similarity measurement methods include but are not limited to some or all of the following:

[0211] (1) Sum of Squared Differences (SSD) algorithm.

[0212] Among them, SSD is a similarity measurement method used to calculate the difference between two image regions. A smaller SSD value indicates that the two regions are more similar.

[0213] In this embodiment of the present application, the SSD values of the pupil feature points extracted from the two iris images can be compared using an SSD algorithm, and corresponding points can be analyzed based on these values. For example, if the SSD value between two pupil feature points is less than a preset SSD threshold, the two pupil feature points can be considered to be a set of corresponding points in the two iris images.

[0214] (2) Normalized Cross Correlation (NCC) algorithm.

[0215] Among them, NCC is another similarity measurement method used to calculate the correlation between two image regions. A larger NCC value indicates that the two regions are more similar.

[0216] In this embodiment of the present application, the NCC values of pupil feature points extracted from the two iris images can be compared using an NCC algorithm, and corresponding points can be identified based on these values. For example, if the NCC value between two pupil feature points is greater than a preset NCC threshold, the two pupil feature points can be considered to be a set of corresponding points in the two iris images.

[0217] It should be noted that the above-mentioned methods for stereo matching of two iris images are only simple examples. In addition, other stereo matching methods are also applicable to the embodiments of the present application and will not be described in detail here.

[0218] S324: Obtain a disparity map corresponding to the two iris images based on the determined positions of the corresponding points in the corresponding iris images.

[0219] Specifically, when the human eye (or camera) observes the same object from two slightly different viewpoints, it sees two slightly different images. This difference between the images is referred to in this article as parallax. Taking the human eye as an example, when we focus our gaze on a distant object, the position of the nearby object in the images seen by the two eyes differs significantly, resulting in a larger parallax. Conversely, the position difference between the images of distant objects in the two eyes is smaller, resulting in a smaller parallax. Based on this principle, our brain is able to interpret these differences (parallax) to understand the three-dimensional structure of the world we see.

[0220] In this context, disparity refers to this spatial difference between images, i.e. the relative position of points in one image in the other. This can be described as a numerical value (e.g., a pixel difference in position) or a brightness level (in a disparity map). This difference, or "disparity," can be used to calculate the depth or distance of an object.

[0221] In an embodiment of the present application, during the process of stereo matching of two iris images, a disparity map is generated, wherein the value of each pixel in the disparity map can represent the horizontal displacement of the corresponding point in the left view and the right view, and a larger value indicates that the object is closer to the observer (such as a binocular camera).

[0222] Assume there is a simplified disparity map, represented as the following two-dimensional matrix:

[0223]

[0224]

[0225] In the disparity map, larger values indicate that the object is closer to the observer.

[0226] This simplified disparity map shows a pupil-like structure. Specifically, in this simplified example, we can see that the value in the middle is 5, the values around it gradually decrease, and the value at the center of the matrix changes the most, indicating that the edge of the pupil may be located in this area.

[0227] In the embodiments of the present application, stereo matching technology is used to more accurately locate the pupil edge under complex lighting and pupil size variations. Accurately finding the pupil edge allows for a more accurate determination of the iris range, improving iris recognition accuracy. This maintains good recognition performance even in complex situations such as pupil size variations, lighting changes, and blurred pupil edges.

[0228] Specifically, after obtaining the disparity map using the stereo matching algorithm, the pupil's depth information can be calculated to determine its position, thereby accurately locating the pupil edge. The specific process is as follows:

[0229] S33: Based on the obtained disparity maps, locate the pupil edge from any one iris image included in the iris image group.

[0230] As described in the above embodiments, the iris image group in the embodiments of the present application may include two iris images or three or more iris images. When the iris image group includes two iris images, one disparity map may be obtained, and when the iris image group includes three or more iris images, multiple disparity maps may be obtained.

[0231] Therefore, when performing step S33 , different positioning methods may be set according to the number of disparity maps.

[0232] Specifically, if a disparity map is obtained, the pupil edge is located from any iris image included in the iris image group directly based on the disparity map.

[0233] If multiple disparity maps are obtained, there are two specific positioning methods:

[0234] Positioning method 1: Disparity map fusion positioning.

[0235] This method is to fuse multiple disparity maps; and then locate the pupil edge from any one iris image included in the iris image group based on the fused disparity map.

[0236] Optionally, when performing image fusion on multiple disparity maps, an optional implementation method is:

[0237] The operation is performed at the pixel level, and the pixel values of corresponding points on different disparity maps are weighted averaged or combined in other forms to obtain the final disparity map.

[0238] like Figure 9 As shown in FIG, which is a schematic diagram of a disparity map fusion in an embodiment of the present application. Assuming there are three simplified disparity maps, each disparity map is represented by a two-dimensional matrix, then Figure 9 The three disparity maps shown can be expressed as:

[0239]

[0240]

[0241]

[0242] In each disparity map, larger values indicate that the object is closer to the observer.

[0243] When performing disparity map fusion, the values of corresponding points in the three disparity maps can be averaged. The final fused disparity map can be expressed as the following two-dimensional matrix:

[0244]

[0245] It should be noted that the above-mentioned disparity map fusion methods are just simple examples. In addition, other image fusion methods are also applicable to the embodiments of the present application and will not be described in detail here.

[0246] Positioning method 2: pupil edge fusion positioning.

[0247] This method is as follows: based on each disparity map, a pupil edge is located from any iris image corresponding to the disparity map; and then, the determined multiple pupil edges are fused to obtain a fused pupil edge.

[0248] Specifically, when performing pupil edge fusion, multiple pupil edges may be aligned in space first, for example, by finding the correspondence between them through feature point matching techniques such as corner detection and descriptor matching.

[0249] An image fusion algorithm can then be applied to merge these edges. Available image fusion methods include, but are not limited to, some or all of the following: Alpha fusion, pyramid fusion, and Poisson fusion. Alpha fusion is a process that superimposes the foreground onto the background through transparency, while pyramid fusion and Poisson fusion offer different image fusion techniques that can blend images while preserving image detail.

[0250] Optionally, the fused pupil edge may be post-processed, such as morphological operations, to remove possible gaps and discontinuous parts to ensure the continuity and integrity of the edge.

[0251] In the above embodiment, whether fusing multiple disparity maps or fusing multiple pupil positioning results to comprehensively determine a pupil edge, the accuracy of pupil positioning can be improved to a certain extent.

[0252] In summary, the above-mentioned methods of locating the pupil edge in multiple views are just simple examples. Other methods are also applicable to the embodiments of the present application and will not be described in detail here.

[0253] In the embodiment of the present application, regardless of which of the above methods is used, the pupil edge can be determined based on a disparity map. Specifically, when locating the pupil edge based on the disparity map, in order to find the position of the pupil edge, the gradient information in the disparity map can be calculated. Among them, the gradient represents the rate of change of the grayscale value in the image, which can be used to detect the edge. Generally, the grayscale change at the edge position is greater. On this basis, after subsequent threshold processing, the position of the pupil edge can be obtained. An optional implementation method is:

[0254] like Figure 10As shown, it is a flow chart of a pupil edge positioning method in an embodiment of the present application. Figure 10 The method shown in FIG. 1 is to locate the pupil edge from the iris image, including the following steps S101 to S103:

[0255] S101: extracting gradient information from the disparity map through an edge detection operator to obtain a gradient map corresponding to the disparity map.

[0256] The gradient element at a certain position in the gradient map represents the rate of change of the grayscale value of the pixel corresponding to the position in the disparity map.

[0257] S102: Determine pupil edge points from the disparity map based on a preset gradient threshold.

[0258] S103: Locate the pupil edge from an iris image according to the determined pupil edge point.

[0259] In an embodiment of the present application, in the disparity map matrix, a slightly larger value (such as the number 5 mentioned above) indicates that the pupil is closer. Because the pupil is sunken in the eyeball, in a relatively stereoscopic image, the pupil area will be farther from the actual camera position in depth of field than the iris (having smaller parallax).

[0260] Next, we need to locate the pupil edge based on the disparity map. To find the edge, this paper uses the method of calculating the grayscale gradient of the disparity map. The gradient is a vector that represents the direction and maximum value of the directional derivative of a function (in this case, the grayscale value of the image) at a specific point. In images, the magnitude of the gradient is often used to represent the strength of the edge.

[0261] Taking the disparity map matrix listed above as an example, the values at the center of the matrix change the most, meaning a significant change from 5 to 3 to 1. The values around it remain relatively stable (all 1). Therefore, the gradient has a larger value in this area. Therefore, the gradient map can be obtained by calculating the difference (i.e., the gradient) between each pixel and its neighbors.

[0262] In the embodiments of the present application, the gradient map calculation process can be implemented using an edge detection operator. The principle of edge detection operators in extracting gradient information is primarily based on the discontinuity of local image characteristics. In an image, edges typically appear as sudden changes in grayscale, color, or texture structure. In the embodiments of the present application, edges can be understood as sudden changes in grayscale in an iris image. Specifically, any one or more of the following edge detection operators can be used to detect these edges.

[0263] Specifically, edge detection operators include but are not limited to some or all of the following:

[0264] Canny operator, Sobel operator, Roberts operator, Prewitt operator, Laplacian operator, LoG operator.

[0265] In general, these operators each have their own advantages and disadvantages, and are suitable for different conditions. For example, the Roberts operator is simple and fast, but sensitive to noise; the Sobel and Prewitt operators are somewhat resistant to noise, but may lose some edge information; the Laplacian operator accurately locates edges, but is very sensitive to noise; and the Canny operator provides a more comprehensive edge detection method, but is computationally more complex. In practical applications, the selection of an appropriate edge detection operator depends on the specific image content and processing requirements, and this article does not provide specific restrictions.

[0266] For example, by extracting the gradient information in the disparity map using the Canny operator, a gradient map corresponding to the disparity map can be obtained.

[0267] Like the disparity map listed above, the gradient map can also be represented as a two-dimensional matrix. The size of this two-dimensional matrix is the same as that of the original image, that is, the gradient map can also be represented as a 5*5 two-dimensional matrix.

[0268] In the embodiment of the present application, after obtaining the gradient map, a threshold is applied to determine which gradient values can be considered edges. This threshold can be fixed or dynamically calculated. By setting the threshold, it is possible to determine which edges are significant and which are likely caused by noise, thereby obtaining the final edge detection result.

[0269] Because a lower threshold value allows more edges to be detected, the result is more susceptible to noise in the image and more likely to pick out irrelevant features from the image. Conversely, a higher threshold value will miss thin or short line segments. Therefore, choosing an appropriate threshold value is important.

[0270] An optional dynamic determination method is:

[0271] In the embodiments of the present application, a threshold selection method with hysteresis can be used to find the pupil edge point using different thresholds. First, an upper threshold is used to find the starting point of the edge. Once a starting point is found, the pupil edge point path is followed point by point on the image. The pupil edge point position is recorded while the value is greater than the lower threshold, and recording stops when the value falls below the lower threshold.

[0272] This method assumes that the pupil edge point is a continuous boundary and is able to follow the blurred part of the pupil edge point seen before detection without marking the noise points in the image as pupil edge points.

[0273] Alternatively, considering that a single global threshold may not be sufficient to process the entire image, we can consider using an adaptive threshold method to adjust the threshold according to the local characteristics of the image.

[0274] Alternatively, you can set a fixed threshold based on experience or experimentation, then filter the gradient array based on the threshold, selecting points with gradient values greater than the threshold as pupil edge points. For example, you can determine the optimal threshold through experimentation. You can start with a low threshold and then gradually increase it to observe the effect of edge detection. The ideal threshold should maximize the extraction of true edges while suppressing noise.

[0275] It should be noted that the two methods of selecting thresholds to find pupil edge points listed above are just simple examples. In addition, other methods are also applicable to the embodiments of the present application and will not be described in detail here.

[0276] In step S103, after determining the pupil edge point, the least squares parabola fitting method (or other fitting methods) can be used to calculate the extreme point coordinates of the edge points in the left and right fixed areas, thereby obtaining the initial center coordinates and radius of the pupil, and then determining the pupil edge.

[0277] In the above embodiment, the present application takes into account that the texture and structural characteristics of the pupil area are different from those of the iris area. If the pupil area is not excluded, it may interfere with the feature extraction. The present application can remove the pupil area from the iris image by locating the pupil edge, thereby reducing interference in the feature extraction process.

[0278] The following describes the subsequent process after determining the pupil edge:

[0279] S34: Determine an iris region in any iris image according to the pupil edge, and perform feature extraction on any iris region to obtain corresponding iris features.

[0280] Combine Figure 2 As can be seen from the diagram of the eye, the iris is an annular area located outside the pupil. Therefore, after determining the pupil edge, the annular area within a certain range of the pupil edge can be used as the iris area. An optional implementation is as follows:

[0281] A preset margin is set around the pupil edge in any iris image; and an annular area determined based on the pupil edge and the preset margin is used as the iris area in any iris image.

[0282] Specifically, in the embodiment of this application, the preset margin represents the difference between the outer and inner diameters of the iris region, i.e., the width of the region. This can generally be determined based on the type of object to be identified, with different preset margins corresponding to different types of objects. For example, the iris region of an adult is approximately 2-4 mm, but this range may vary from person to person; for example, a preset margin of 3 mm may be set. The iris region of an infant is generally narrower, approximately 1-2 mm; for example, a preset margin of 1.5 mm may be set.

[0283] For example, the width of the iris area of cats and dogs varies depending on breed, age, and individual differences. Generally speaking, the width of a cat's iris area is about 1-2 mm, so the preset margin can be set to 1.5 mm, for example; while the width of a dog's iris area is about 2-4 mm, so the preset margin can be set to 3 mm, etc.

[0284] It should be noted that the preset margin values listed above are just simple examples. They can be flexibly set due to individual differences, etc., and this article does not make specific restrictions.

[0285] like Figure 11 As shown in FIG, it is a schematic diagram of a method for determining the iris area in an embodiment of the present application. The black circular area is the pupil, and the white circle in the black circular area represents the pupil edge. When the pupil edge and the preset margin r are determined, the annular area outside the pupil edge with a width of the preset margin r can be determined as the iris area. Figure 11 Medium gray area.

[0286] In the above embodiment, the position and size information of the pupil edge can help determine the range of the iris. Specifically, by setting a certain margin around the pupil edge, an iris area that is unrelated to the pupil can be obtained. Since this area contains the main texture information of the iris, it is beneficial to improve the accuracy of feature extraction.

[0287] After determining the iris area based on the above method, it is necessary to extract the texture features of the iris. The main purpose is to extract discriminative features from the iris image that can reflect its texture and structural information. These features should have high discrimination and stability so as to be used in the subsequent recognition and matching process.

[0288] In the embodiment of the present application, feature extraction is performed on the iris region to obtain corresponding iris features, including at least one of the following methods:

[0289] Feature extraction method 1: Extract features of the iris area through filters of different scales and directions to obtain corresponding response values; combine the response values extracted by each filter to form the iris features corresponding to the iris area.

[0290] For example, a Gabor filter is used to extract texture features of the iris region. The Gabor filter is a method commonly used in texture analysis and feature extraction. It can capture local texture information of an image at different scales and directions.

[0291] The direction of the Gabor filter is determined by an angle parameter θ defined by the filter kernel. This angle parameter describes the direction of the parallel stripes in the kernel. In practical applications, the direction parameter can take any real value between 0° and 360°, which allows the Gabor filter to respond to features in different directions in the image. This directional selectivity makes the Gabor filter particularly suitable for processing texture information and has been widely used in vision science because it can simulate the sensitivity of the human visual system to directional features.

[0292] The scale of the Gabor filter is a parameter related to the frequency bandwidth and directional selectivity of the Gabor filter. The scale parameter is usually related to the standard deviation σ of the Gaussian envelope function, which determines the width of the filter in the frequency domain. A larger σ value means that the filter has a wider response range in the frequency domain, thereby capturing more frequency components; while a smaller σ value corresponds to a narrower frequency domain response, only capturing frequency components within a specific range. By adjusting these parameters, the Gabor filter can be designed to respond only to image features of a specific scale, making it very useful in multi-scale analysis.

[0293] In iris recognition in this application, Gabor filters can effectively extract iris texture features. Specifically, a series of Gabor filters of different scales and orientations are applied to the iris image. The Gabor filter responses capture the rich details of the iris texture based on spectral and spatial local characteristics. The extracted response values can be used to represent the texture information in the iris image, and these response values are combined to form the feature vector of the iris.

[0294] Among them, the selection of parameters such as Gabor filter scale and direction needs to adapt to the characteristics of iris texture, and an adaptive method can be used to select appropriate filter parameters.

[0295] In the above embodiment, the Gabor filter can simulate the response of the human visual system and effectively extract iris features. Moreover, by adjusting the filter parameters, it can be better adapted to the texture characteristics of the iris, thereby improving the accuracy and robustness of recognition.

[0296] It should be noted that, in addition to the above-mentioned method of extracting iris features using Gabor filters, other filters may also be used to extract iris features, such as Gaussian filters, Butterworth filters, Laplace filters, etc., which will not be detailed here.

[0297] In addition, when selecting filters, it is necessary to consider the characteristics of iris texture and the needs of subsequent feature matching and classification. It is also possible to combine multiple filters and analysis methods to extract more discriminative iris texture features, such as combining Gabor filters and Gaussian filters.

[0298] Feature extraction method 2: By comparing the grayscale value of each pixel in the iris area with the corresponding neighboring pixel points, the iris features corresponding to the iris area are extracted.

[0299] This method is called Local Binary Pattern (LBP). LBP is a texture description method that can describe local texture features in an image. Specifically, the LBP algorithm compares the grayscale values of a pixel with those of its neighbors and generates a binary sequence as the LBP value for that pixel.

[0300] In iris recognition in this application, LBP can be used to extract local texture features of the iris. Applying LBP to the entire iris image can obtain a feature vector describing the iris texture.

[0301] In the above implementation, both the Gabor filter and the LBP algorithm can effectively capture iris texture features, thereby providing strong support for subsequent recognition and matching processes.

[0302] It should be noted that the feature extraction methods listed above are just simple examples. In actual applications, other methods can also be used to extract iris features, such as those based on the LoG operator, scale-invariant feature transform (SIFT), accelerated robust feature extraction (SURF), Fourier transform, multi-scale analysis methods, etc., which can effectively extract the feature information of iris texture to improve the accuracy and robustness of iris recognition. They will not be listed here one by one.

[0303] The Fourier transform converts an image from the spatial domain to the frequency domain, enabling processing with frequency-domain filters. By analyzing the spectral characteristics of iris texture, corresponding frequency-domain filters can be designed to extract specific texture features. Multiscale analysis methods, such as wavelet transform and multiresolution analysis, can also be used to extract iris texture features. These methods analyze images at different scales, thereby capturing the multiscale characteristics of iris texture.

[0304] Any of the feature extraction methods listed above can be used alone or in combination with other feature extraction methods, which is not specifically limited in this article.

[0305] It should be noted that the texture and structural features of the pupil area are different from those of the iris area. In the embodiment of the present application, by locating the pupil edge, the pupil area can be removed from the iris image, eliminating the interference of the pupil area in feature extraction, thereby reducing interference in the iris feature extraction process.

[0306] In an embodiment of the present application, considering that the texture information of the iris area near the edge of the pupil may be affected by lighting, occlusion, etc., by locating the edge of the pupil, corresponding processing strategies can be adopted for these affected areas, such as adjusting filter parameters, enhancing contrast, etc., to optimize the feature extraction effect.

[0307] An optional implementation is to enhance the contrast of the iris region before extracting features from the iris region to obtain corresponding iris features.

[0308] For example, the iris region may be subjected to histogram equalization processing again, or local contrast enhancement, adaptive contrast enhancement, gamma correction, or filtering may be performed to enhance the contrast of the iris region.

[0309] It should be noted that any method of enhancing image contrast is applicable to the embodiments of the present application and will not be described in detail here.

[0310] Another optional implementation is that before extracting features from the iris region and obtaining corresponding iris features, at least one disparity map may be processed in the following manner to update the iris region. For the specific process, see Figure 12 ,like Figure 12 As shown, it is a flowchart of an iris area updating method in an embodiment of the present application, including the following steps S121 to S124:

[0311] S121: Extracting gradient information in the disparity map through different edge detection operators to obtain multiple gradient maps corresponding to the disparity map.

[0312] Among them, each edge detection operator corresponds to a gradient map; the gradient elements in the gradient map represent: the rate of change of the grayscale value of each pixel in the disparity map.

[0313] S122: Determine a new pupil edge point from the disparity map based on the multiple gradient maps.

[0314] In step S121, for a disparity map, a plurality of different edge detection operators may be used to perform gradient calculation, for example, using the Sobel operator, Roberts operator, Prewitt operator, LoG operator, and Canny operator to process the disparity map. Thereafter, step S122 may be performed.

[0315] Specifically, when multiple edge detection operators are used simultaneously, the gradient maps extracted by each edge detection operator can be fused. For example, the gradient amplitudes obtained by different operators can be weighted averaged. On this basis, a threshold is applied to the fused gradient map to find a new pupil edge point. The threshold in this case can be a fixed threshold previously set based on experience or experiments; or it can be a new threshold obtained by adjusting the fixed threshold based on the fused gradient map and the characteristics of different edge detection operators; or it can be a threshold selected by using the threshold selection method with hysteresis or the adaptive threshold method listed above. For details, please refer to the above embodiments and will not be repeated here.

[0316] The method of fusing the gradient map is the same as the method of fusing the disparity map, and will not be described in detail here.

[0317] S123: Update the iris region in any iris image according to the new pupil edge point.

[0318] Similar to S103 above, in step S123, after determining the new pupil edge point, the least squares parabola fitting method (or other fitting methods) can be used to calculate the extreme point coordinates of the edge points in the left and right fixed areas, thereby obtaining the initial center coordinates and radius of the pupil, and then determining the new pupil edge.

[0319] On this basis, a preset margin can be set around the new pupil edge in any iris image, and the annular area determined based on the new pupil edge and the preset margin can be used as the new iris area. After extracting the new iris area, the iris features of the area can be re-extracted. The specific implementation method can be found in the above embodiment, and the repeated parts will not be repeated.

[0320] Through the above implementation, the extraction of iris features can be optimized. Even when texture information may be affected by illumination, occlusion, etc., more accurate iris features can be effectively extracted to improve the accuracy of subsequent recognition.

[0321] Specifically, after the iris features are extracted, they need to be compared with the iris features stored in the pre-built database to achieve individual identity recognition. The specific process is as follows:

[0322] S35: Based on the iris features, the identity of the object to be identified is identified.

[0323] In the embodiment of the present application, the recognition process generally adopts a feature matching algorithm, such as calculating the cosine distance, Hamming distance, Minkowski distance, etc. between two iris features to measure the similarity between the two iris features, thereby determining whether they belong to the same individual.

[0324] Taking the calculation of the cosine distance between two iris features as an example, the iris features in the embodiment of the present application can be expressed in the form of feature vectors. The cosine similarity between the two feature vectors is determined. This can be achieved by calculating the dot product of the two vectors and then dividing it by the product of their respective moduli. The obtained cosine similarity can then be converted into a cosine distance. It is assumed that the obtained cosine distance is a decimal between 0 and 1, where 0 represents exactly the same direction and 1 represents completely opposite directions.

[0325] If the cosine distance between two iris features is less than a certain threshold, the two iris features can be considered to match. Then, the identity information of the object to be identified can be determined based on the identity information corresponding to the matched iris features.

[0326] like Figure 13 As shown in FIG, it is a schematic diagram of an object identification process by feature matching in an embodiment of the present application. Assume that there are 4 iris features stored in the database, such as Figure 13 As shown, the iris features of the object to be identified need to be matched with the iris features in the database respectively. If the second feature is matched successfully, the identity information of the object to be identified can be determined based on the identity information associated with each iris feature in the database. Figure 13 The identification result of the object to be identified includes name, gender, age, etc.

[0327] It should be noted that the databases and identity information listed above are just simple examples. In actual applications, the number of iris features stored in the database can be more or less, and accordingly, the stored identity information can also be simpler or more complex. The above are just simple examples and this article does not make specific limitations.

[0328] Based on the above, this application takes into account that over-reliance on a certain type of image may ignore useful information contained in other types of images. In order to further improve the accuracy of iris recognition, multi-source image information, such as infrared images, visible light images, etc., can be combined to improve the accuracy of pupil edge positioning, thereby improving the accuracy of iris recognition.

[0329] That is, in step S31, iris images may be collected in multiple image modes. Iris images under multiple collection viewing angles may be collected in each image mode, thereby obtaining multiple iris image groups, each iris image group corresponding to one image mode.

[0330] In the embodiment of the present application, different image modes refer to different ways of capturing and representing image data.

[0331] For example, in this application, the following image modes are included but not limited to:

[0332] RGB image mode, infrared image mode, visible light image mode.

[0333] Among them, the iris image collected in the RGB image mode is an RGB image. This type of image is relatively common and will not be described in detail here.

[0334] The iris image collected in the infrared image mode is an infrared image. Since the pupil has a high transmittance to infrared light, the infrared image can highlight the pupil area.

[0335] The iris image collected in the visible light image mode is a visible light image, and the visible light image can clearly show the structure of the entire eye.

[0336] First, it is necessary to collect both infrared and visible light images of the pupil. This can be done with specialized equipment, such as a camera with both infrared and visible light modes.

[0337] Then, before step S32, iris images in multiple image modes should be fused. An optional implementation is as follows:

[0338] First, multiple iris image groups are divided into candidate iris image sets corresponding to different acquisition perspectives.

[0339] The candidate iris images in each candidate iris image set are iris images collected in different image modes and at the same collection viewing angle.

[0340] like Figure 14 As shown, it is a logical diagram of the division of a candidate iris image set in an embodiment of the present application. Figure 14 The iris images collected by the binocular camera in three image modes are listed. The images collected by the two cameras of the binocular camera can be recorded as the left view and the right view. Then, in the three image modes of RGB, infrared and visible light, three iris image groups are collected, such as Figure 14 Iris image group 1 includes a left RGB image and a right RGB image, iris image group 2 includes a left infrared image and a right infrared image, and iris image group 3 includes a left visible light image and a right visible light image.

[0341] In this embodiment of the present application, the number of candidate iris image sets is consistent with the number of acquisition perspectives, and the number of candidate iris images in a candidate iris image set is consistent with the number of image modes. That is, iris images in the same candidate iris image set are iris images acquired in different image modes and at the same acquisition perspective.

[0342] like Figure 14In the image, the binocular camera corresponds to two acquisition perspectives, and two candidate iris image sets can be obtained. For example, candidate iris image set 1 includes: left RGB image, left infrared image and left visible light image; candidate iris image set 2 includes: right RGB image, right infrared image and right visible light image.

[0343] Then, for each pixel point, the corresponding pixel points on each candidate iris image in the same candidate iris image set are subjected to feature fusion to obtain a fused iris image.

[0344] In this process, an optional implementation method for the feature fusion between images is to perform a weighted average of the grayscale values of corresponding pixels on each candidate iris image in the same candidate iris image set.

[0345] Specifically, the grayscale values of the pixels at the same location in multiple candidate iris images are weighted averaged to obtain the grayscale value of the pixel at that location in the corresponding fused iris image. The weights used in this weighted average can be flexibly set based on actual needs or determined through experimentation. For example, when fusing RGB, infrared, and visible light images, the weights for each can be set to 1 / 3. This is not a specific limitation in this article.

[0346] Optionally, for subsequent pupil edge localization, it may be necessary to place greater emphasis on information in the infrared image, as the infrared image can better highlight the pupil. Therefore, when fusing the RGB, infrared, and visible light images, instead of using a 1 / 3 weighting for each, the infrared image can be weighted more highly than the other two. For example, the weights of the RGB, infrared, and visible light images can be set to 1 / 4, 1 / 2, and 1 / 4, respectively. Another example is setting the weights of the RGB, infrared, and visible light images to 1 / 6, 1 / 2, and 1 / 3, respectively.

[0347] In addition to the grayscale value, other pixel features can also be fused, such as color, transparency, brightness, saturation, etc., which will not be repeated here.

[0348] It should also be noted that in order to further improve the accuracy of iris recognition, in addition to fusing multi-source images, the above-listed methods can also be used to pre-process each iris image in each iris image group, such as denoising and histogram equalization of the collected images to improve image quality. The specific implementation methods can be found in the above embodiments and will not be repeated here.

[0349] Based on the fused iris image, when executing step S32, stereo matching is performed between each two fused iris images to obtain a corresponding disparity map. Then, subsequent processes such as pupil edge positioning, iris region determination, and iris feature extraction can be performed. The specific implementation method can be found in the above embodiment, and the repeated parts are not repeated here.

[0350] In the embodiment of the present application, richer and more accurate information can be obtained by fusing multiple source images. On this basis, when the subsequent recognition process is performed through the fused image, recognition can be performed based on richer and more accurate information, further improving the accuracy of iris recognition.

[0351] The following modularized description of the specific process of iris recognition is given by taking the use of a binocular camera to collect an iris image group containing two iris images (also called an iris image pair) as an example:

[0352] See Figure 15 , which is a modular block diagram of an iris recognition process in an embodiment of the present application. Specifically, the above process can be simply divided into the following modules: image acquisition module, preprocessing module, stereo matching module, feature extraction module, recognition and matching module.

[0353] Among them, the image acquisition module can collect iris images of the same eye area of the object to be identified based on the binocular camera, and obtain Figure 15 The iris image pair shown.

[0354] Then, each iris image in the iris image pair can be preprocessed by a preprocessing module to improve image quality. Specifically, operations such as denoising and histogram equalization can be performed on the iris image to improve image quality. Figure 15 In this article, histogram equalization is taken as an example. First, the grayscale histogram of the image needs to be statistically analyzed, and on this basis, the cumulative distribution function of each grayscale level is calculated. Finally, the cumulative distribution function is used to map the grayscale value of the original image to the new grayscale value, so that the grayscale distribution of the new image is more uniform, completing the histogram equalization process.

[0355] After histogram equalization processing, the contrast of the image is improved and the iris texture is clearer, which is beneficial to the subsequent pupil edge positioning and feature extraction operations.

[0356] In the stereo matching module, a stereo matching algorithm can be used to accurately locate the pupil edge. The process can be summarized as follows: obtaining an iris image pair; correcting the images: Before stereo matching, the images need to be corrected to eliminate the parallax between the left and right eye images; then performing feature extraction. In the corrected images, pupil feature points are extracted for matching. After extracting the feature points, a stereo matching algorithm can be used to find corresponding points in the left and right eye images and obtain a disparity map. After obtaining the disparity map, the pupil edge can be located by calculating its depth information. For detailed implementation, please refer to the above embodiment, and any repetitions will not be repeated here.

[0357] In the feature extraction module, based on Figure 15 One or more of the methods listed above are used to extract iris features for subsequent recognition and matching processes. For specific implementation methods, please refer to the above embodiments, and repeated details will not be repeated.

[0358] In the recognition and matching module, the extracted iris features can be matched with the iris features in the database to obtain identity recognition results. The specific implementation method can be found in the above embodiment, and the repeated parts will not be repeated.

[0359] In summary, the iris recognition method of the embodiment of the present application can effectively reduce pupil edge positioning errors and improve iris recognition accuracy. Even in complex situations such as pupil size changes, light changes, and blurred pupil edges, it still has good recognition results.

[0360] Based on the same inventive concept, the present application also provides an iris recognition device. Figure 16 As shown in FIG, it is a schematic diagram of the structure of the iris recognition device 1600, which may include:

[0361] The image acquisition unit 1601 is configured to acquire an iris image group collected for the object to be identified. The iris image group includes: multiple iris images collected from different viewing angles for the same eye region of the object to be identified;

[0362] The stereo matching unit 1602 is configured to perform stereo matching between each pair of iris images in the iris image group to obtain a corresponding disparity map, wherein the disparity elements in each disparity map represent the displacement of corresponding points between the two iris images in a specified direction.

[0363] a pupil locating unit 1603 configured to locate a pupil edge from any one of the iris images included in the iris image group based on the obtained disparity maps;

[0364] An iris feature extraction unit 1604 is configured to determine an iris region in any iris image based on the pupil edge, and perform feature extraction on any iris region to obtain corresponding iris features;

[0365] The identification unit 1605 is configured to perform identity recognition on the subject based on iris features.

[0366] Optionally, the stereo matching unit 1602 is specifically configured to:

[0367] Eliminating the parallax between the two iris images by correcting the two iris images;

[0368] Extract pupil feature points from each rectified iris image respectively;

[0369] Through stereo matching, corresponding points are determined from the extracted pupil feature points;

[0370] Based on the determined positions of the corresponding points in the corresponding iris images, a disparity map corresponding to the two iris images is obtained.

[0371] Optionally, the pupil locating unit 1603 is specifically configured to:

[0372] If a disparity map is obtained, then based on the disparity map, the pupil edge is located from any one iris image included in the iris image group;

[0373] If multiple disparity maps are obtained, the multiple disparity maps are fused, and based on the fused disparity map, the pupil edge is located from any iris image included in the iris image group; alternatively, based on each disparity map, a pupil edge is located from any iris image corresponding to the disparity map, and the multiple determined pupil edges are fused to obtain a fused pupil edge.

[0374] Optionally, the pupil locating unit 1603 is specifically configured to:

[0375] Based on a disparity map, the pupil edge is located from an iris image as follows:

[0376] The gradient information in the disparity map is extracted through the edge detection operator to obtain the gradient map corresponding to the disparity map; the gradient elements in the gradient map represent: the rate of change of the grayscale value of each pixel in the disparity map;

[0377] Based on a preset gradient threshold, the pupil edge point is determined from the disparity map;

[0378] The pupil edge is located from an iris image according to the determined pupil edge point.

[0379] Optionally, the iris feature extraction unit 1604 is specifically configured to:

[0380] Set a preset margin around the pupil edge in any iris image;

[0381] The annular area determined based on the pupil edge and the preset margin is used as the iris area in any iris image.

[0382] Optionally, the iris feature extraction unit 1604 is specifically configured to perform at least one of the following steps:

[0383] The iris region is subjected to feature extraction through filters of different scales and directions to obtain corresponding response values; the response values extracted by each filter are combined to form the iris features corresponding to the iris region;

[0384] By comparing the grayscale values of each pixel in the iris area with the corresponding neighboring pixels, the iris features corresponding to the iris area are extracted.

[0385] Optionally, the iris feature extraction unit 1604 is further configured to enhance the contrast of the iris region before performing feature extraction on the iris region to obtain corresponding iris features.

[0386] Optionally, the iris feature extraction unit 1604 is further configured to process at least one disparity map in the following manner to update the iris region before extracting features from the iris region to obtain corresponding iris features:

[0387] Different edge detection operators are used to extract gradient information from the disparity map, resulting in multiple gradient maps corresponding to the disparity map. Each edge detection operator corresponds to a gradient map. The gradient elements in the gradient map represent the rate of change of the grayscale value of each pixel in the disparity map.

[0388] Based on multiple gradient maps, new pupil edge points are determined from the disparity map;

[0389] According to the new pupil edge point, the iris region in any iris image is updated.

[0390] Optionally, if there are multiple iris image groups, and each iris image group corresponds to an image pattern, the stereo matching unit 1602 is further configured to:

[0391] Dividing the multiple iris image groups into candidate iris image sets corresponding to different acquisition perspectives, wherein the candidate iris images in each candidate iris image set are iris images acquired in different image modes and at the same acquisition perspective;

[0392] For each pixel point, feature fusion is performed on the corresponding pixel points on each candidate iris image in the same candidate iris image set to obtain a fused iris image;

[0393] The stereo matching unit 1602 is specifically used for:

[0394] Stereo matching is performed between each two fused iris images to obtain the corresponding disparity map.

[0395] Optionally, the image mode includes some or all of the following:

[0396] RGB image mode, infrared image mode, visible light image mode.

[0397] Optionally, the stereo matching unit 1602 is specifically configured to:

[0398] The grayscale values of the corresponding pixels on each candidate iris image in the same candidate iris image set are weighted averaged.

[0399] Optionally, before performing stereo matching between each two iris images to obtain corresponding disparity maps, the stereo matching unit 1602 is further configured to:

[0400] For each iris image in the iris image group, image enhancement processing is performed on the iris image in at least one manner.

[0401] For the convenience of description, the above parts are divided into modules (or units) according to their functions and described separately. Of course, when implementing this application, the functions of each module (or unit) can be implemented in the same or two software or hardware.

[0402] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program that has a predetermined function and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories) or a combination thereof. Similarly, a processor (or two processors or memories) can be used to implement one or two modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0403] After introducing the iris recognition method and apparatus according to an exemplary embodiment of the present application, an electronic device according to another exemplary embodiment of the present application will be introduced next.

[0404] Those skilled in the art will appreciate that various aspects of the present application can be implemented as systems, methods, or program products. Therefore, various aspects of the present application can be specifically implemented in the following forms: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation that combines hardware and software aspects, which may be collectively referred to herein as a "circuit," "module," or "system."

[0405] Based on the same inventive concept as the above method embodiment, an electronic device is also provided in the embodiment of the present application. In one embodiment, the electronic device may be a server. In this embodiment, the structure of the electronic device may be as follows: Figure 17 As shown, it includes a memory 1701 , a communication module 1703 and one or two processors 1702 .

[0406] Memory 1701 is used to store computer programs executed by processor 1702. Memory 1701 may mainly include a program storage area and a data storage area. The program storage area may store an operating system and programs required for running instant messaging functions, while the data storage area may store various instant messaging messages and operating instruction sets.

[0407] Memory 1701 may be a volatile memory, such as random-access memory (RAM); a non-volatile memory, such as read-only memory, flash memory, a hard disk drive (HDD), or a solid-state drive (SSD); or any other medium capable of carrying or storing a desired computer program in the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory 1701 may be a combination of the aforementioned memories.

[0408] The processor 1702 may include one or two central processing units (CPUs) or a digital processing unit, etc. The processor 1702 is configured to implement the above iris recognition method when calling the computer program stored in the memory 1701 .

[0409] The communication module 1703 is used to communicate with terminal devices and other servers.

[0410] The specific connection medium between the memory 1701, the communication module 1703 and the processor 1702 is not limited in the embodiment of the present application. Figure 17 In the embodiment, the memory 1701 and the processor 1702 are connected via a bus 1704. Figure 17 The connections between the other components are shown in bold lines for illustration only and are not intended to be limiting. The bus 1704 can be divided into an address bus, a data bus, a control bus, etc. For ease of description, Figure 17 The diagram shows a single thick line, but this does not indicate that there is only one bus or one type of bus.

[0411] The memory 1701 stores a computer storage medium, which stores computer executable instructions. The computer executable instructions are used to implement the iris recognition method of the embodiment of the present application. The processor 1702 is used to execute the above iris recognition method, such as Figure 3 shown.

[0412] In another embodiment, the electronic device may also be other electronic devices, such as terminal devices. In this embodiment, the structure of the electronic device may be as follows: Figure 18 As shown, it includes: a communication component 1810, a memory 1820, a display unit 1830, a camera 1840, a sensor 1850, an audio circuit 1860, a Bluetooth module 1870, a processor 1880 and other components.

[0413] The communication component 1810 is used to communicate with the server. In some embodiments, it may include a wireless fidelity (WiFi) module. The WiFi module is a short-range wireless transmission technology. Electronic devices can help users send and receive information through the WiFi module.

[0414] Memory 1820 can be used to store software programs and data. Processor 1880 executes various functions and data processing of terminal device 110 by running the software programs or data stored in memory 1820. Memory 1820 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Memory 1820 stores an operating system that enables terminal device 110 to operate. In the present application, memory 1820 can store the operating system and various application programs, and may also store computer programs that execute the iris recognition method according to the embodiments of the present application.

[0415] The display unit 1830 can also be used to display information input by the user or provided to the user, as well as a graphical user interface (GUI) for various menus of the terminal device 110. Specifically, the display unit 1830 may include a display screen 1832 disposed on the front of the terminal device 110. The display screen 1832 may be configured in the form of a liquid crystal display, a light-emitting diode, etc. The display unit 1830 can be used to display the identity recognition results, etc. in the embodiments of the present application.

[0416] The display unit 1830 can also be used to receive input digital or character information and generate signal input related to the user settings and function control of the terminal device 110. Specifically, the display unit 1830 may include a touch screen 1831 arranged on the front of the terminal device 110, which can collect user touch operations on or near it, such as clicking a button, dragging a scroll box, etc.

[0417] The touch screen 1831 can be covered on the display screen 1832, or the touch screen 1831 and the display screen 1832 can be integrated to realize the input and output functions of the terminal device 110. The integrated display screen can be simply called a touch screen. In this application, the display unit 1830 can display applications and corresponding operation steps.

[0418] Camera 1840 can be used to capture still images, and users can publish images captured by camera 1840 through an application. There can be one or two cameras 1840. The object generates an optical image through the lens and projects it onto the photosensitive element. The photosensitive element can be a charge coupled device (CCD) or a complementary metal-oxide-semiconductor (CMOS) phototransistor. The photosensitive element converts the optical signal into an electrical signal, which is then transmitted to processor 1880 for conversion into a digital image signal.

[0419] The terminal device may further include at least one sensor 1850, such as an acceleration sensor 1851, a distance sensor 1852, a fingerprint sensor 1853, and a temperature sensor 1854. The terminal device may also be configured with other sensors such as a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, a light sensor, and a motion sensor.

[0420] The audio circuit 1860, speaker 1861, and microphone 1862 provide an audio interface between the user and the terminal device 110. The audio circuit 1860 can convert the received audio data into an electrical signal and transmit it to the speaker 1861, which converts it into a sound signal for output. The terminal device 110 may also be equipped with a volume button for adjusting the volume of the sound signal. On the other hand, the microphone 1862 converts the collected sound signal into an electrical signal, which is received by the audio circuit 1860 and converted into audio data. The audio data is then output to the communication component 1810 for transmission to, for example, another terminal device 110, or the audio data is output to the memory 1820 for further processing.

[0421] The Bluetooth module 1870 is used to exchange information with other Bluetooth devices having a Bluetooth module through the Bluetooth protocol. For example, the terminal device can establish a Bluetooth connection with a wearable electronic device (such as a smart watch) that also has a Bluetooth module through the Bluetooth module 1870 to exchange data.

[0422] Processor 1880 is the control center of the terminal device, connecting various parts of the entire terminal using various interfaces and lines. By running or executing software programs stored in memory 1820 and calling data stored in memory 1820, it performs various functions of the terminal device and processes data. In some embodiments, processor 1880 may include one or two processing units; processor 1880 may also integrate an application processor and a baseband processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the baseband processor mainly processes wireless communications. It is understood that the above-mentioned baseband processor may not be integrated into processor 1880. In this application, processor 1880 can run the operating system, application programs, user interface display and touch response, as well as the iris recognition method of the embodiment of this application. In addition, processor 1880 is coupled to display unit 1830.

[0423] In some possible implementations, various aspects of the iris recognition method provided in the present application may also be implemented in the form of a program product, which includes a computer program. When the program product is run on an electronic device, the computer program is used to enable the electronic device to perform the steps of the iris recognition method according to various exemplary embodiments of the present application described above in this specification. For example, the electronic device may perform the following steps: Figure 3 Follow the steps shown in .

[0424] The program product may employ any combination of one or two readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or two conductors, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0425] The program product of the embodiment of the present application may be a portable compact disc read-only memory (CD-ROM) and include a computer program, and can be run on an electronic device. However, the program product of the present application is not limited thereto. In this document, a readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with a command execution system, apparatus, or device.

[0426] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a readable computer program. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with a command execution system, apparatus, or device.

[0427] The computer program embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0428] The computer program for performing the operations of the present application can be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and conventional procedural programming languages such as "C" or similar programming languages. The computer program can be executed entirely on the user electronic device, partially on the user electronic device, as a separate software package, partially on the user electronic device and partially on a remote electronic device, or entirely on a remote electronic device or server. In cases involving remote electronic devices, the remote electronic device can be connected to the user electronic device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external electronic device (for example, using an Internet service provider to connect through the Internet).

[0429] It should be noted that although several units or subunits of the device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, depending on the embodiment of the application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by two units.

[0430] Furthermore, although the operations of the method of the present application are described in a particular order in the accompanying drawings, this does not require or imply that the operations must be performed in this particular order, or that all illustrated operations must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, two steps may be combined into one step, and / or one step may be decomposed into two steps.

[0431] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or two computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing a computer-usable computer program.

[0432] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program commands. These computer program commands can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the commands executed by the processor of the computer or other programmable data processing device generate commands for implementing the processes in the flowchart and / or block diagram. Figure 1 One or two processes and / or boxes Figure 1 A device that has the functions specified in one or both boxes.

[0433] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising a command device, which implements the process Figure 1 One or two processes and / or boxes Figure 1 The function specified in one or both boxes.

[0434] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 One or two processes and / or boxes Figure 1 The steps for the function specified in one or both boxes.

[0435] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0436] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. An iris recognition method, characterized in that: The method comprises: Acquire an iris image group collected for the object to be identified; the iris image group includes: multiple iris images collected at different acquisition viewing angles for the same eye area of the object to be identified; In the iris image group, stereo matching is performed between each two iris images to obtain corresponding disparity maps, wherein the disparity elements in each disparity map represent: the displacement of each corresponding point between the two corresponding iris images in a specified direction; Locating a pupil edge from any one iris image included in the iris image group based on the obtained disparity maps; Determining an iris region in any one of the iris images according to the pupil edge, and performing feature extraction on the any one of the iris regions to obtain corresponding iris features; Based on the iris features, the identity of the object to be identified is identified.

2. The method according to claim 1, wherein The stereo matching is performed between each two iris images to obtain a corresponding disparity map, including: Eliminating the parallax between the two iris images by correcting the two iris images; Extract pupil feature points from each rectified iris image respectively; Through stereo matching, corresponding points are determined from the extracted pupil feature points; Based on the determined positions of the corresponding points in the corresponding iris images, a disparity map corresponding to the two iris images is obtained.

3. The method according to claim 1, wherein The step of locating a pupil edge from any one of the iris images included in the iris image group based on the obtained disparity maps includes: If a disparity map is obtained, locating a pupil edge from any one iris image included in the iris image group based on the disparity map; If multiple disparity maps are obtained, the multiple disparity maps are image fused, and based on the fused disparity map, the pupil edge is located from any iris image included in the iris image group; alternatively, based on each disparity map, a pupil edge is located from any iris image corresponding to the disparity map, and the multiple determined pupil edges are fused to obtain a fused pupil edge.

4. The method according to any one of claims 1 to 3, wherein: Based on a disparity map, the pupil edge is located from an iris image as follows: Extracting gradient information from the disparity map using an edge detection operator to obtain a gradient map corresponding to the disparity map; a gradient element in the gradient map represents: a rate of change of the grayscale value of each pixel in the disparity map; Determining a pupil edge point from the disparity map based on a preset gradient threshold; The pupil edge is located from the one iris image according to the determined pupil edge point.

5. The method according to any one of claims 1 to 3, wherein: The determining, based on the pupil edge, an iris region in any one of the iris images includes: Setting a preset margin around the pupil edge in any one of the iris images; An annular area determined based on the pupil edge and the preset margin is used as the iris area in any one iris image.

6. The method according to any one of claims 1 to 3, wherein: The extracting features of the iris region to obtain corresponding iris features includes at least one of the following methods: Extracting features from the iris region using filters of different scales and directions to obtain corresponding response values; combining the response values extracted by each filter to form iris features corresponding to the iris region; By comparing the grayscale values of each pixel point in the iris area with the corresponding neighboring pixel points, the iris features corresponding to the iris area are extracted.

7. The method according to claim 6, wherein Before extracting features from the iris area to obtain corresponding iris features, the method further includes: The contrast of the iris region is enhanced.

8. The method according to claim 6, wherein Before extracting features from the iris area to obtain corresponding iris features, the method further includes: Processing at least one disparity map to update the iris region is performed in the following manner: Extracting gradient information from the disparity map using different edge detection operators to obtain multiple gradient maps corresponding to the disparity map; wherein each edge detection operator corresponds to a gradient map; and gradient elements in the gradient map represent: a rate of change of the grayscale value of each pixel in the disparity map; determining a new pupil edge point from the disparity map based on the multiple gradient maps; The iris region in any one iris image is updated according to the new pupil edge point.

9. The method according to any one of claims 1 to 3, wherein: If there are multiple iris image groups, and each iris image group corresponds to an image pattern, then before performing stereo matching between each two iris images to obtain corresponding disparity maps, the method further includes: Dividing the plurality of iris image groups into candidate iris image sets corresponding to different acquisition perspectives, wherein the candidate iris images in each candidate iris image set are iris images acquired in different image modes and at the same acquisition perspective; For each pixel point, feature fusion is performed on the corresponding pixel points on each candidate iris image in the same candidate iris image set to obtain a fused iris image; The stereo matching is performed between each two iris images to obtain a corresponding disparity map, including: Stereo matching is performed between each two fused iris images to obtain the corresponding disparity map.

10. The method according to claim 9, wherein The image mode includes some or all of the following: RGB image mode, infrared image mode, visible light image mode.

11. The method according to claim 9, wherein The step of fusing features of corresponding pixels on each candidate iris image in the same candidate iris image set includes: The grayscale values of the corresponding pixels on each candidate iris image in the same candidate iris image set are weighted averaged.

12. The method according to any one of claims 1 to 3, wherein: Before performing stereo matching between each two iris images to obtain corresponding disparity maps, the method further includes: For each iris image in the iris image group, image enhancement processing is performed on the iris image in at least one manner.

13. An iris recognition device, characterized in that: include: An image acquisition unit is configured to acquire an iris image group collected for an object to be identified; the iris image group comprises: a plurality of iris images collected at different acquisition viewing angles for the same eye region of the object to be identified; a stereo matching unit, configured to perform stereo matching between each two iris images in the iris image group to obtain a corresponding disparity map, wherein the disparity elements in each disparity map represent the displacement of corresponding points between the two iris images in a specified direction; a pupil locating unit, configured to locate a pupil edge from any one iris image included in the iris image group based on the obtained disparity maps; an iris feature extraction unit, configured to determine an iris region in any one iris image according to the pupil edge, and perform feature extraction on the any one iris region to obtain corresponding iris features; The identification unit is used to identify the object to be identified based on the iris feature.

14. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor is enabled to perform the steps of any one of the methods of claims 1 to 12.

15. A computer-readable storage medium, characterized in that The method comprises a computer program. When the computer program is run on an electronic device, the computer program is used to enable the electronic device to execute the steps of any one of the methods according to claims 1 to 12.

16. A computer program product, characterized in that The method comprises a computer program stored in a computer-readable storage medium; when a processor of an electronic device reads the computer program from the computer-readable storage medium, the processor executes the computer program, so that the electronic device performs the steps of any one of the methods described in claims 1 to 12.

Images