An eye movement tracking method and system compatible with iris recognition

Abstract

The present invention relates to the technical field of VR recognition, and specifically discloses an eye movement tracking method and system compatible with iris recognition. The method includes: confirming that the current task is one of iris recognition, eye movement tracking, or a combination of both, and collecting the first human eye image or the second human eye image at different resolutions when the tasks are different; calculating the eye line of sight from the first human eye image; calculating the segmentation masks of the iris annular region and the pupil region from the second human eye image, detecting the reflection spots, calculating the center coordinates of all the spots and their corresponding relationships with the infrared lamps, and identifying the invalid spots; turning off the infrared lamps corresponding to the invalid spots; performing feature encoding on the iris texture in the continuously collected second human eye image, obtaining the iris texture features and then comparing them to obtain the similarity of different irises, and performing recognition based on the similarity. The present invention can solve the problem of incompatibility between iris recognition and eye movement tracking in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of VR recognition, and particularly to an eye movement tracking method and system compatible with iris recognition. Background Art

[0002] Iris recognition is an important biometric recognition technology means. It determines a person's identity by comparing the similarity between iris image features and is a recognition technology with significant security performance in multimodal biometrics (second only to DNA recognition technology). The processing process of an iris recognition system generally involves collecting an eye image through an infrared camera, and then locating the iris region through image processing and analysis methods, and encoding and comparing the iris texture.

[0003] Eye movement tracking is a type of technology that tracks eye movement, line of sight direction, and line of sight landing point. It has gradually become a standard technology for current XR form products such as smart glasses, VR / AR headsets, etc. Such devices have high requirements for portability. If an iris recognition function is added to such products, the reuse of hardware for iris recognition and eye movement tracking needs to be considered. The commonality of these two technologies is that generally an infrared camera is required to collect an eye image for analysis and processing. The difference is that the iris recognition task requires obtaining a clear and undisturbed iris image as much as possible, so it has high requirements for the resolution of the collected image and tries to ensure that the iris region is not disturbed by the reflection spots of the infrared lamp. The eye movement tracking task has little requirement for the clarity of the iris image texture, emphasizes the processing efficiency of the system more, and at the same time, the most mainstream eye movement tracking method, the pupil-corneal reflection method, needs to ensure that multiple stable reflection spots can be obtained when the pupil stares at different directions. Therefore, if eye movement tracking is implemented according to the standard of iris recognition, the tracking efficiency will be low; if iris recognition is implemented according to the standard of eye movement tracking, the accuracy will be low.

[0004] Therefore, implementing an eye movement tracking method and system that can be compatible with iris recognition has significant practical significance. Summary of the Invention

[0005] In view of the above technical problems, the present invention provides an eye movement tracking method and system compatible with iris recognition to solve the problem of incompatibility between iris recognition and eye movement tracking in the prior art.

[0006] Other features and advantages of the present disclosure will become apparent through the following detailed description, or be learned in part through the practice of the present disclosure.

[0007] According to one aspect of the present invention, an eye movement tracking method compatible with iris recognition is disclosed. The method includes:

[0008] Confirm that the current task is one of iris recognition, eye movement tracking, or a combination of iris recognition and eye movement tracking. When the user is performing eye movement tracking, drive the infrared camera to collect the first human eye image illuminated by the infrared lamp at the first resolution. When the user is performing only iris recognition or iris recognition and eye movement tracking simultaneously, drive the infrared camera to switch to the second resolution to collect the second human eye image illuminated by multiple infrared lamps. The positional relationship between the multiple infrared lamps and the infrared camera satisfies that at least one reflection spot is within the pupil region range of the second human eye image, and the second resolution is greater than the first resolution;

[0009] When the first human eye image is collected, calculate the user's eye line of sight from the first human eye image;

[0010] When the second human eye image is collected, perform iris detection and iris segmentation on the second human eye image to obtain the segmentation masks of the iris annular region and the pupil region of the human eye image;

[0011] Detect the reflection spots in the iris annular region and the pupil region, obtain the central coordinates of all the reflection spots and their corresponding relationships with the infrared lamps, and after registration, obtain the reflection spots that are invalid for the iris recognition task;

[0012] When performing the iris recognition task, control and turn off the infrared lamps corresponding to the reflection spots that are invalid for the iris recognition task;

[0013] Perform feature encoding on the iris texture in the iris annular region of the continuously collected second human eye image to obtain iris texture features;

[0014] Compare different iris texture features to obtain the similarity of different irises, and identify the user's identity based on the similarity.

[0015] Further, the calculating the user's eye line of sight from the first human eye image includes:

[0016] Collect the pupil image of the user's same eye, and based on the CNN network, calculate the bounding box of the pupil to detect the pupil center;

[0017] Locate the pupil center of the user based on the reflection spots in the pupil image, and use the vector line from the center of the eyeball model to the pupil center as the user's eye line of sight. The eyeball model is established in advance, and when the eyeball model is established, it includes:

[0018] Continuously capture the pupil images of the same eye of the user, project the infrared camera to the pupil center corresponding thereto into three-dimensional space based on the multi-view geometry principle and the internal and external parameters of the infrared camera to obtain rays, and use the intersection points or the closest points of different rays as the three-dimensional positions of the pupil center. Fit the multiple three-dimensional positions of the pupil center and combine the empirical constant of the eyeball radius to obtain the eyeball model of the user.

[0019] Further, when performing iris detection and iris segmentation on the second eye image, it includes:

[0020] Divide the second eye image into blocks according to a preset rule to obtain multiple sub-images;

[0021] Based on edge detection and morphological operations, initially identify the iris edge in the sub-image, and based on the graphic fitting algorithm, initially locate the pupil boundary in the sub-image;

[0022] Based on the local region growth algorithm, segment and extract the regions of the iris and pupil initially detected in each sub-image, and based on edge correction and morphological processing, segment the sub-images with regions of the iris and pupil;

[0023] Fuse the regions of the iris and pupil extracted from each sub-image to form a complete segmentation mask of the iris annular region and the pupil region.

[0024] Further, when fusing, the method further includes:

[0025] When there are discontinuities or conflicts between two adjacent sub-images, perform integration based on the interpolation or boundary smoothing algorithm;

[0026] After fusion, the method further includes:

[0027] Perform edge smoothing and small region removal on the global contours of the reconstructed iris annular region and pupil region;

[0028] Verify whether the shape, size, and position of the reconstructed iris annular region and pupil region meet the physiological standards.

[0029] Further, when registering the reflection spot, it includes:

[0030] Judge the distance from the reflection spot to the pupil region, and discard the reflection spots not within the threshold range;

[0031] Judge the shape of the reflection spot, and discard the reflection spots whose shapes do not conform to the preset rules;

[0032] Fit the distribution shape of the reflected light spots, and filter out the reflected light spots whose fitting rules do not correspond to the distribution rules of the infrared lamps.

[0033] Further, a plurality of the infrared lamps and the infrared camera are concentrically distributed on the same plane.

[0034] Further, when detecting the reflected light spots in the iris annular region and the pupil region, it specifically includes:

[0035] Based on the edge detection algorithm, calculate the edges of the respective reflected light spots in the iris annular region and the pupil region to obtain the shapes of the respective reflected light spots;

[0036] According to the positions of the edge pixels of the respective reflected light spots, obtain the average value to get the central coordinates;

[0037] According to the position layout of the infrared lamps, compare the relative position relationships of the central coordinates of the respective reflected light spots to obtain the corresponding relationships between the respective central coordinates and the plurality of infrared lamps.

[0038] Further, when performing feature encoding on the iris texture, use the rubbersheet transform combined with the 2D-Gabor transform to perform feature encoding on the iris texture.

[0039] Further, when obtaining the similarity of different irises through comparison, use the Hamming distance to measure the similarity.

[0040] According to the second aspect of the present disclosure, there is provided an eye movement tracking device compatible with iris recognition, including: an acquisition and resolution switching module, which confirms that the current task is one of iris recognition, eye movement tracking, or a combination of iris recognition and eye movement tracking. When the user performs eye movement tracking, it drives the infrared camera to acquire the first human eye image illuminated by the infrared lamp at the first resolution. When the user performs only iris recognition or iris recognition and eye movement tracking simultaneously, it drives the infrared camera to switch to the second resolution to acquire the second human eye image illuminated by the plurality of infrared lamps. The positional relationship between the plurality of infrared lamps and the infrared camera satisfies that at least one reflected light spot is within the pupil region range of the second human eye image, and the second resolution is greater than the first resolution;

[0041] An eye movement tracking module, which is used to calculate the user's eye line of sight from the first human eye image when the first human eye image is acquired;

[0042] An iris positioning module, which is used to perform iris detection and iris segmentation on the second human eye image when the second human eye image is acquired to obtain the segmentation masks of the iris annular region and the pupil region of the human eye image;

[0043] A spot detection module, configured to detect reflected light spots within the iris annular region and the pupil region, obtain the central coordinates of all the reflected light spots and their corresponding relationships with the infrared lamps, and after registration, obtain reflected light spots that are invalid for the iris recognition task;

[0044] A lighting control module, configured to control and turn off the infrared lamps corresponding to the reflected light spots that are invalid for the iris recognition task when performing the iris recognition task;

[0045] A feature encoding module, configured to perform feature encoding on the iris texture within the iris annular region of the continuously acquired second human eye image to obtain iris texture features;

[0046] A feature comparison module, configured to compare different iris texture features to obtain the similarity of different irises, and identify the identity of the user based on the similarity.

[0047] The technical solution of the present disclosure has the following beneficial effects:

[0048] By combining eye movement tracking and iris recognition, and through resolution switching, eye movement tracking can be achieved at a low resolution, and iris recognition can be achieved at a high resolution. Therefore, the high efficiency of eye movement tracking and the high accuracy of iris recognition are maintained, and the compatibility of the device is greatly improved. Description of the Drawings

[0049] Figure 1 It is a flowchart of an eye movement tracking method compatible with iris recognition in an embodiment of this specification;

[0050] Figure 2 It is a structural block diagram of an eye movement tracking device compatible with iris recognition in an embodiment of this specification;

[0051] Figure 3 It is a position structure diagram of an infrared camera and an infrared lamp in an embodiment of this specification;

[0052] Figure 4 It is a schematic diagram of a reflected light spot within the pupil region range in an embodiment of this specification. Detailed Embodiments

[0053] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of the present disclosure. However, those skilled in the art will realize that the technical solutions of the present disclosure can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be used. In other cases, well-known technical solutions are not shown or described in detail to avoid obscuring the various aspects of the present disclosure.

[0054] In addition, the accompanying drawings are only schematic illustrations of the present disclosure. The same reference numerals in the drawings denote the same or similar parts, and thus the repeated description thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0055] As Figure 1 shown, the embodiments of the present specification provide an eye movement tracking method compatible with iris recognition. The execution subject of this method can be a computer, a server, a wearable intelligent device, etc. This method can specifically include the following steps S101 to S104:

[0056] In step S101, it is confirmed that the current task is one of iris recognition, eye movement tracking, or a combination of iris recognition and eye movement tracking. When the user performs eye movement tracking, the infrared camera is driven to collect a first human eye image illuminated by an infrared lamp at a first resolution. When the user performs only iris recognition or iris recognition and eye movement tracking are performed simultaneously, the infrared camera is driven to switch to a second resolution to collect a second human eye image illuminated by a plurality of the infrared lamps. The positional relationship between the plurality of infrared lamps and the infrared camera satisfies that at least one reflected spot is within the pupil area range of the second human eye image, and the second resolution is greater than the first resolution.

[0057] Specifically, first, confirm whether the current task type is iris recognition, eye movement tracking, or a combination of both. When the system detects that the user is performing eye movement tracking, the infrared camera is set to capture images at a lower first resolution, which are captured after being illuminated by an infrared light. This can improve the speed and efficiency of image acquisition and meet the real-time requirements of eye movement tracking. If it is detected that the task is only iris recognition or both iris recognition and eye movement tracking, the camera will switch to a higher second resolution mode for image acquisition. These images are also illuminated by an infrared light. In this case, multiple infrared lights are used, and the positions of these lights are properly arranged with respect to the camera. When the user is looking straight ahead for iris recognition, preferably two reflection spots are within the pupil area of the captured image. This ensures that high-quality and high-detail images can be captured during iris recognition, thereby improving the accuracy of recognition. Therefore, the second resolution is higher than the first resolution to meet the requirements for image clarity in iris recognition. Among them, the second resolution is not lower than 640x480, and the first resolution is lower than 640x480.

[0058] Exemplarily, the positional structure of the infrared camera and the infrared light is as Figure 3 shown. When the user is looking straight ahead for iris recognition, two reflection spots are within the pupil area of the captured image, as Figure 4 shown. To make the reflection spots prominent, Figure 4 the reflection spots within the pupil area are marked with yellow circles, and the reflection spots within the iris area are marked with red circles.

[0059] In step S102, when the first human eye image is captured, calculate the user's eye line of sight from the first human eye image.

[0060] Among them, calculating the user's eye line of sight from the first human eye image includes: capturing the pupil image of the same eye of the user, calculating the bounding box of the pupil based on the CNN network to detect the pupil center; positioning the pupil center of the user based on the reflection spots in the pupil image, and using the vector line from the center of the eyeball model to the pupil center as the user's eye line of sight. The eyeball model is established in advance. When establishing the eyeball model, it includes: continuously capturing the pupil images of the same eye of the user, projecting the infrared camera to the corresponding pupil center into three-dimensional space based on the principles of multi-view geometry and the internal and external parameters of the infrared camera to obtain rays, taking the intersection point or the closest point of different rays as the three-dimensional position of the pupil center, fitting multiple three-dimensional positions of the pupil center, and combining the empirical constant of the eyeball radius to obtain the user's eyeball model.

[0061] For explanation purposes, pupil images are collected from the user's eyes, and these images focus on the pupil area. During the image processing process, the system uses a Convolutional Neural Network (CNN) to identify the bounding box of the pupil. CNN is a deep learning model commonly used for image recognition, which can automatically extract features from images and identify the boundaries of the pupil through these features. By identifying the boundaries of the pupil, the central position of the pupil can then be further calculated. After obtaining the initial position of the pupil center, the reflection spots in the image are further used to accurately locate the pupil center. The reflection spots are bright spots formed by the infrared light source on the cornea. Due to the reflection characteristics of the cornea, the positions of these spots have a direct geometric relationship with the position of the pupil. By analyzing the positions of these spots, the position of the pupil center can be determined more accurately. Additionally, in order to accurately calculate the line of sight direction, it is necessary to establish a user's eye model in advance. The process of establishing this model involves capturing multiple pupil images of the same eye of the user and using the principles of multi-view geometry to relate these images to the physical relationships in three-dimensional space. Specifically, the principles of multi-view geometry are used to derive the position of an object in three-dimensional space by observing the same object from multiple different perspectives. In this process, the internal parameters (such as focal length, principal point coordinates) and external parameters (such as the rotation and translation parameters of the camera) of the infrared camera are used to determine the specific position and direction of the camera when capturing images. Through these parameters, the system can project the relationship between the camera and the corresponding pupil center into three-dimensional space to form rays, and then analyze these rays captured from different perspectives to find the intersection point or the closest point between the rays, and determine this point as the position of the pupil center in three-dimensional space. Since the positions of the camera and the pupil may change over time, the system continuously captures multiple images and extracts multiple three-dimensional position points from these images. Subsequently, the system fits these three-dimensional position points and combines the known radius of the eyeball (this radius is usually based on empirical data, such as 12MM) to finally generate a complete eye model. After obtaining the eye model, after the subsequent iris recognition identifies the user's identity, the eye model of this user can be called. In this way, in subsequent eye movement tracking, only by locating the coordinates of the user's pupil center according to the spots, eye movement tracking can be quickly carried out, that is, the vector line between the center point of the eye model and the pupil center point is regarded as the user's line of sight direction. Since the eye model has been established in three-dimensional space, the direction of the line of sight can be derived through simple geometric calculations.

[0062] In step S103, when the second human eye image is collected, iris detection and iris segmentation are performed on the second human eye image to obtain the segmentation masks of the iris annular region and the pupil region of the human eye image.

[0063] Specifically, when performing iris detection and iris segmentation on the second human eye image, it includes: dividing the second human eye image into blocks according to a preset rule to obtain multiple sub-images; initially identifying the iris edge in the sub-images based on edge detection and morphological operations, and initially locating the pupil boundary in the sub-images based on a graphic fitting algorithm; segmenting and extracting the initially detected iris and pupil regions of each sub-image based on a local region growth algorithm, and segmenting the sub-images with regions having iris and pupil based on edge correction and morphological processing; fusing the extracted iris and pupil regions in each sub-image to form a complete segmentation mask for the iris annular region and the pupil region.

[0064] In the above, the second human eye image block division can be grid block division, that is, dividing the image into multiple equal-sized blocks (such as 8x8 or 16x16 grids) according to a predetermined grid size. Each block is processed separately, which is convenient for parallel computing; it can also be key region priority block division, that is, according to prior knowledge or simple feature detection (such as edge detection or pupil detection), the pupil region and the iris edge region are preferentially divided, and then the remaining part is divided into grids. In each block, lightweight edge detection or morphological operations are applied to quickly and initially identify the iris edge. For the block containing the pupil region, a simple circular fitting algorithm (such as the Hough transform) is further used to accurately locate the pupil boundary. After initial identification, local segmentation is performed on the initially detected iris region in each block, and a threshold-based method (such as the Otsu method) or a local region growth algorithm is used to extract the iris and pupil regions. For the block of the iris edge region, detailed edge correction and morphological processing (such as dilation and erosion) are performed to ensure the accuracy of the segmentation boundary. Finally, the iris and pupil boundary data extracted from each block are fused to form a complete segmentation mask for the iris annular region and the pupil region.

[0065] In step S104, the reflection light spots in the iris annular region and the pupil region are detected to obtain the central coordinates of all the reflection light spots and their correspondence with the infrared lamp, and after registration, the reflection light spots invalid for the iris recognition task are obtained.

[0066] Specifically, when detecting the reflected light spots in the iris annular area and the pupil area, the method specifically includes: calculating the edges of each of the reflected light spots in the iris annular area and the pupil area based on an edge detection algorithm (such as the Candy algorithm) to obtain the shape of each of the reflected light spots; obtaining the center coordinates by taking the average value according to the position of the edge pixels of each of the reflected light spots; and comparing the relative positional relationship of the center coordinates of each of the reflected light spots according to the position layout of the infrared lamps to obtain the corresponding relationship between each of the center coordinates and the plurality of infrared lamps. After that, the registration can be performed to confirm the invalid reflected light spots.

[0067] When the reflected light spot is registered, one or more of the following criteria are included: pupil distance criterion, light spot shape criterion, and light spot distribution criterion.

[0068] Pupil distance criterion: judging the distance from the reflected light spot to the pupil area, and discarding the reflected light spot that is not within the threshold range; Figure 3 As shown in the figure, the infrared lamp and the infrared camera are distributed in a circle, so they can be placed in the AR helmet. However, because the relative change range of the distance between the AR helmet camera and the eye is not too large, when the eyeball moves, the change of the pupil center is also within a certain range. Therefore, based on the pupil distance criterion, some ineffective corneal reflection spots can be effectively filtered out.

[0069] Spot shape criterion: The shape of the reflected light spot is judged, and the reflected light spot whose shape does not conform to the preset rule is discarded; the shape of the effective reflected light spot is basically close to a circle and has a high aggregation reading. The light spots with obviously abnormal aspect ratio and low aggregation degree are generally the spots with divergence and tailing phenomena at the boundary of the iris and sclera. Therefore, these spots can be discarded.

[0070] Light spot distribution criterion: fit the distribution shape of the reflected light spot, and filter out the reflected light spot whose fitting rule does not correspond to the distribution rule of the infrared lamp, that is, the distribution of the effective reflected light spot should correspond to the distribution of the infrared lighting source, fit multiple light spot distribution shapes, and filter out points that obviously do not conform to the fitting rule.

[0071] Furthermore, the plurality of infrared lamps and the infrared camera are distributed on the same plane and concentrically.

[0072] In step S105, when the iris recognition task is performed, the infrared lights corresponding to the reflection light spots that are invalid for the iris recognition task are controlled to be turned off.

[0073] Among them, according to the result output of step S104, the infrared lamp is controlled by the infrared lamp control unit, and only the reflection light spots registered as valid above are maintained. Generally speaking, after registration, only the infrared lamps with reflection light spots within the pupil area are kept on, and the rest of the infrared lamps are turned off to ensure that the texture of the iris area in the image is not interfered by the reflection light spots.

[0074] In step S106, feature encoding is performed on the iris texture within the iris annular region of the continuously acquired second human eye image to obtain iris texture features.

[0075] As a supplement, when performing feature encoding on the iris texture, the rubbersheet transform combined with the 2D-Gabor transform is used to perform feature encoding on the iris texture.

[0076] Among them, since the structure of the iris is usually affected by factors such as pupil dilation, contraction, and eyeball rotation during capture, resulting in image deformation. To overcome these problems, the rubbersheet transform converts the iris annular region from the polar coordinate system to the rectangular coordinate system, unfolds the circular iris into a rectangular planar image, which is equivalent to flattening the annular region of the iris, and can normalize it to a fixed size, eliminating the deformation effects caused by pupil size changes or eyeball rotation, ensuring that under different acquisition conditions, the iris texture features can maintain consistency, thereby improving the accuracy and stability of subsequent feature extraction. Next, on the iris annular region normalized by the rubbersheet transform, the 2D-Gabor transform is used for feature encoding.

[0077] The 2D-Gabor transform is a spatial frequency-based filtering technique that can effectively extract directional and local texture information in an image. The response of the Gabor filter can well simulate the receptive field characteristics in the human visual system and is particularly suitable for extracting textures with periodic or directional features. By filtering the iris image with a set of Gabor filters with different directions and scales, the spatial frequency information of the iris texture in multiple directions can be captured, that is, the Gabor transform will generate a set of filter responses with different phases and directions, and these response values reflect the distribution characteristics of the iris texture in each direction and scale. During the feature encoding process, the output of each Gabor filter forms a texture feature vector, and these vectors are combined to form the final iris texture feature. Since the Gabor filter can effectively capture the fine structural information in the iris, such as stripes, spots, etc., the encoded feature vector can very accurately represent an individual's iris features.

[0078] By combining the rubber sheet transformation and the 2D-Gabor transformation, the influence of image deformation can be effectively eliminated, and highly distinguishable iris texture features can be extracted. These features not only have high stability and consistency under different shooting conditions, but also possess strong discrimination ability, and can be used for accurate authentication and identification in the iris recognition system.

[0079] In step S107, different iris texture features are compared to obtain the similarity of different irises, and the identity of the user is recognized based on the similarity.

[0080] Among them, when obtaining the similarity of different irises through comparison, the similarity is measured based on the Hamming distance. The Hamming distance is a measurement method commonly used for calculating the similarity between binary codes, and it quantifies the difference between two binary strings by calculating the differences in the corresponding bits. For iris recognition, each iris image will be encoded as a fixed-length binary feature vector after processing such as rubber sheet transformation and 2D-Gabor transformation, and these vectors can be represented as strings composed of 0s and 1s. The process of calculating the Hamming distance is to compare two iris feature vectors bit by bit. If the bit values at the corresponding positions are different, it is counted as one difference.

[0081] Specifically, first, the iris image to be compared is transformed into a binary feature vector through rubber sheet transformation and 2D-Gabor transformation. Then, all the binary feature vectors in the database are called one by one, and these vectors are compared bit by bit. The total number of differences in the bit values at all corresponding positions is counted, and the result obtained is the Hamming distance. Next, according to a pre-set threshold, it is judged whether the similarity is high enough to confirm the identity. If the Hamming distance is less than the threshold, it is considered that the two irises come from the same person, thus confirming the user's identity; otherwise, it is considered that the two irises belong to different people, and the identity authentication request is rejected.

[0082] In an embodiment, when fusing the regions of the iris and the pupil extracted from each of the subgraphs to form a complete segmentation mask of the iris annular region and the pupil region, it specifically includes:

[0083] When there is discontinuity or conflict between two adjacent subgraphs, integration is performed based on the interpolation or boundary smoothing algorithm.

[0084] And, after fusing the regions of the iris and the pupil, perform:

[0085] Perform edge smoothing and small region removal on the global contours of the reconstructed iris annular region and the pupil region;

[0086] Verify whether the shapes, sizes, and positions of the reconstructed iris annular region and the pupil region conform to physiological standards.

[0087] Based on the same idea, as Figure 2 shown, an exemplary embodiment of the present disclosure also provides an eye movement tracking system compatible with iris recognition, including:

[0088] An acquisition and resolution switching module 201, which confirms that the current task is one of iris recognition, eye movement tracking, or a combination of iris recognition and eye movement tracking. When the user performs eye movement tracking, it drives the infrared camera to acquire the first human eye image illuminated by the infrared lamp at the first resolution. When the user performs only iris recognition or iris recognition and eye movement tracking simultaneously, it drives the infrared camera to switch to the second resolution to acquire the second human eye image illuminated by multiple infrared lamps. The positional relationship between the multiple infrared lamps and the infrared camera is such that at least one reflected light spot is within the pupil region range of the second human eye image, and the second resolution is greater than the first resolution;

[0089] An eye movement tracking module 202, which is used to calculate the user's eye line of sight for the first human eye image when the first human eye image is acquired.

[0090] An iris positioning module 203, which is used to perform iris detection and iris segmentation on the second human eye image when the second human eye image is acquired, to obtain the segmentation masks of the iris annular region and the pupil region of the human eye image.

[0091] A light spot detection module 204, which is used to detect the reflected light spots within the iris annular region and the pupil region, obtain the central coordinates of all the reflected light spots and their corresponding relationships with the infrared lamps, and after registration, obtain the reflected light spots that are invalid for the iris recognition task.

[0092] A lighting control module 205, which is used to control and turn off the infrared lamps corresponding to the reflected light spots that are invalid for the iris recognition task when performing the iris recognition task.

[0093] A feature encoding module 206, which is used to perform feature encoding on the iris texture within the iris annular region of the continuously acquired second human eye image to obtain iris texture features.

[0094] A feature comparison module 207, which is used to compare different iris texture features to obtain the similarity of different irises, and identify the user's identity based on the similarity.

[0095] The above system combines eye movement tracking with iris recognition. Through resolution switching, it can achieve eye movement tracking at a low resolution and iris recognition at a high resolution. Therefore, it maintains the high efficiency of eye movement tracking and the high accuracy of iris recognition, greatly improving the compatibility of the device.

[0096] From the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described herein can be implemented by software or by a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (such as a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the exemplary embodiments of the present disclosure.

[0097] In addition, the above drawings are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, rather than for limiting purposes. It is easy to understand that the processes shown in the above drawings do not indicate or limit the chronological order of these processes. Additionally, it is also easy to understand that these processes can be executed synchronously or asynchronously in, for example, multiple modules.

[0098] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the claims.

[0099] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. An eye movement tracking method compatible with iris recognition, characterized in that, The method includes: Confirm that the current task is one of iris recognition, eye movement tracking, or a combination of iris recognition and eye movement tracking. When the user performs eye movement tracking, drive the infrared camera to collect the first human eye image illuminated by the infrared lamp at the first resolution. When the user performs only iris recognition or iris recognition and eye movement tracking simultaneously, drive the infrared camera to switch to the second resolution to collect the second human eye image illuminated by multiple infrared lamps. The positional relationship between the multiple infrared lamps and the infrared camera is such that at least one reflected spot is within the pupil region of the second human eye image, and the second resolution is greater than the first resolution. When the first human eye image is collected, calculate the user's eye line of sight from the first human eye image. When the second human eye image is collected, perform iris detection and iris segmentation on the second human eye image to obtain the segmentation masks of the iris annular region and the pupil region of the human eye image. Detect the reflected spots in the iris annular region and the pupil region, obtain the central coordinates of all the reflected spots and their corresponding relationships with the infrared lamps, and after registration, obtain the reflected spots that are invalid for the iris recognition task. When performing the iris recognition task, control and turn off the infrared lamps corresponding to the reflected spots that are invalid for the iris recognition task. Perform feature encoding on the iris texture in the iris annular region of the continuously collected second human eye image to obtain iris texture features. Compare different iris texture features to obtain the similarity of different irises, and identify the user's identity based on the similarity.

2. The eye movement tracking method compatible with iris recognition according to claim 1, wherein When performing iris detection and iris segmentation on the second human eye image, it includes: Divide the second human eye image into blocks according to a preset rule to obtain multiple sub-images. Based on edge detection and morphological operations, initially identify the iris edge in the sub-image, and based on the graphic fitting algorithm, initially locate the pupil boundary in the sub-image. Based on the local region growth algorithm, segment and extract the regions of the iris and pupil initially detected in each sub-image, and based on edge correction and morphological processing, segment the sub-images with regions of the iris and pupil. Fuse the regions of the iris and pupil extracted from each sub-image to form the segmentation masks of the complete iris annular region and the pupil region.

3. The eye movement tracking method compatible with iris recognition according to claim 2, wherein, When fusing, the method further includes: When there are discontinuities or conflicts between two adjacent sub-images, perform integration based on the interpolation or boundary smoothing algorithm. After fusion, the method further includes: Perform edge smoothing and small region removal on the global contours of the reconstructed iris annular region and pupil region. Verify whether the shapes, sizes, and positions of the reconstructed iris annular region and pupil region meet the physiological standards.

4. The eye movement tracking method compatible with iris recognition according to claim 1, wherein When registering the reflected spots, it includes: Judge the distance from the reflected spots to the pupil region, and discard the reflected spots outside the threshold range. Judge the shape of the reflected spots, and discard the reflected spots whose shapes do not conform to the preset rules. The distribution shape of the reflected light spots is fitted, and the reflected light spots whose fitting rules do not correspond to the distribution rules of the infrared lamps are filtered out.

5. The eye movement tracking method compatible with iris recognition according to claim 1, wherein The plurality of infrared lamps and the infrared camera are distributed concentrically on the same plane.

6. The eye movement tracking method compatible with iris recognition according to claim 1, wherein When detecting the reflected light spots in the iris annular area and the pupil area, the method specifically includes: Calculate the edges of each of the reflection light spots in the iris annular area and the pupil area based on an edge detection algorithm to obtain the shape of each of the reflection light spots; According to the positions of the edge pixels of each of the reflection spots, an average value is calculated to obtain the center coordinates; According to the position layout of the infrared lamps, the relative position relationship of the center coordinates of each of the reflected light spots is compared to obtain the corresponding relationship between each of the center coordinates and the plurality of infrared lamps.

7. The eye movement tracking method compatible with iris recognition according to claim 1, characterized in that When encoding the features of iris texture, rubbersheet transform combined with 2D-Gabor transform is used to encode the features of iris texture.

8. The eye movement tracking method compatible with iris recognition according to claim 1, wherein, When the similarity of different irises is obtained by comparison, the similarity is measured based on the Hamming distance.

9. An eye movement tracking system compatible with iris recognition, characterized in that, include: an acquisition and resolution switching module, which confirms that the current task is one of iris recognition, eye tracking, or a combination of iris recognition and eye tracking, and when the user performs eye tracking, drives the infrared camera to acquire a first human eye image illuminated by an infrared lamp at a first resolution; when the user performs only iris recognition or iris recognition and eye tracking at the same time, drives the infrared camera to switch to a second resolution to acquire a second human eye image illuminated by a plurality of the infrared lamps, and the positional relationship between the plurality of the infrared lamps and the infrared camera satisfies that at least one reflected light spot is located within the pupil area of ​​the second human eye image, and the second resolution is greater than the first resolution; an eye tracking module, for calculating the user's eye sight line based on the first human eye image when the first human eye image is collected; an iris positioning module, configured to perform iris detection and iris segmentation on the second human eye image when the second human eye image is collected, so as to obtain segmentation masks of the iris annular area and the pupil area of ​​the human eye image; A light spot detection module is used to detect the reflected light spots in the iris annular area and the pupil area, obtain the center coordinates of all the reflected light spots and their corresponding relationship with the infrared light, and obtain the reflected light spots that are invalid for the iris recognition task after registration; A light control module, used for controlling and turning off the infrared light corresponding to the reflected light spot that is invalid for the iris recognition task when performing the iris recognition task; A feature encoding module, used for performing feature encoding on the iris texture in the iris annular region in the second human eye image that is continuously collected, to obtain iris texture features; The feature comparison module is used to compare the different iris texture features to obtain the similarity of different irises and identify the identity of the user based on the similarity.

Images