Strabismus pupil positioning method, device, computer-readable storage medium and device

Abstract

The present invention discloses a strabismus pupil location method, apparatus, readable storage medium, and device, belonging to the field of iris recognition. The method comprises: initially locating the pupil boundary as a circle on an iris image; using the initial location result as a reference, dividing the iris image into a plurality of sub-images according to a certain center angle; for each sub-image, traversing a center search range and a radius search range, and using the center and radius values ​​corresponding to the maximum grayscale change as the center and radius of the sub-image; obtaining an arc segment for each sub-image based on the center, radius, and center angle of each sub-image, and splicing the arc segments of all sub-images together to obtain the pupil boundary. The present invention improves the accuracy of strabismus pupil location, and ultimately improves the accuracy of iris recognition.

Description

Technical Field

[0001] The present invention relates to the field of iris recognition, and in particular to a strabismus pupil positioning method, device, computer-readable storage medium and equipment. Background Art

[0002] Biometrics technology uses computers, optics, acoustics, biosensors, and biostatistics to authenticate individuals using inherent physiological characteristics (such as fingerprints, faces, and irises) and behavioral traits (such as handwriting, voice, and gait). Iris recognition is a form of biometric identification technology. As a key identification feature, the iris offers advantages such as lifelong uniqueness, stability, collectibility, and non-invasiveness. It is an inevitable trend in the development of identity recognition research and applications.

[0003] The iris is a ring-like structure located between the pupil and the sclera. Figure 1 The outer and inner irises of the middle iris lose some iris information due to obstruction by the eyelids and eyelashes. The iris is approximately 12 mm in diameter and 0.5 mm thick. From an identification perspective, the iris's uniqueness is reflected by its interwoven subtle features, resembling filaments and stripes. These features are typically considered iris texture features and are used for iris recognition.

[0004] Iris recognition mainly includes iris image acquisition, iris image quality assessment, iris image preprocessing, iris image normalization, iris image feature extraction and iris feature comparison.

[0005] In iris recognition, due to equipment experience, lack of cooperation from the data collector and other external factors, sometimes the pupil and iris of the collected iris image may be distorted, such as strabismus. Figure 2 In iris recognition, it is necessary to accurately locate the pupil boundary.

[0006] Currently, there are two commonly used pupil boundary positioning methods for iris recognition:

[0007] 1) Treat the pupil as a circle. First, perform gradient processing and sharpening on the image. Common image sharpening methods include the Sobel algorithm and the Canny operator. This sharpening method filters the image, sets a threshold, and highlights the pupil boundary. Then, using this boundary point information, curve fitting methods such as symmetric radial transform, least squares method, and Hough transform are used to accurately locate the pupil boundary as a circle.

[0008] This method is for Figure 2 The strabismus pupil image may have a portion of its boundary fitting inaccurate because the boundary of the strabismus pupil is not in the shape of a circle.

[0009] 2) Directly perform point-to-point alignment according to the pupil boundary. First, use a method similar to 1) to highlight the pupil boundary. Then, use the pupil boundary as a set of point arrays and arrange these points counterclockwise or clockwise to depict the pupil boundary.

[0010] Although this method breaks free from the limitation of a circle, noise such as eyelashes and light spots may occasionally obscure some boundary points of the pupil. In other words, this set of pupil boundary point arrays cannot describe some boundary points, resulting in omissions. This has a significant impact on the subsequent expansion, normalization, and feature extraction processes. Summary of the Invention

[0011] In order to solve the defect of inaccurate positioning of the existing strabismus pupil positioning method, the present invention provides a strabismus pupil positioning method, device, readable storage medium and equipment, which improve the accuracy of strabismus pupil positioning and ultimately improve the accuracy of iris recognition.

[0012] The present invention provides the following technical solutions:

[0013] In a first aspect, the present invention provides a method for locating a strabismus pupil, the method comprising:

[0014] The pupil boundary is initially positioned as a circle on the iris image to obtain the initial positioning circle center and initial positioning radius;

[0015] Taking the initially positioned circle center as the center, the iris image is divided into a plurality of sub-images according to a certain central angle;

[0016] Based on the initial positioning center and initial positioning radius, set the center search range and radius search range;

[0017] For each sub-image, traverse the center search range and radius search range, and take the center value and radius value corresponding to the maximum grayscale change as the center and radius of the sub-image;

[0018] The arc segment of each sub-image is obtained according to the center, radius and central angle of each sub-image, and the arc segments of all sub-images are spliced ​​together to obtain the pupil boundary.

[0019] Furthermore, for each sub-image, traversing the circle center search range and the radius search range, and taking the circle center value and radius value corresponding to the maximum grayscale change as the circle center and radius of the sub-image, includes:

[0020] For each sub-image, expand the annular region image corresponding to the sub-image with the center value (x, y) and the radius search range [r1, r2] into a rectangular region image;

[0021] Constructing a filter of size n*1, and performing a convolution operation on the filter and the rectangular area image to obtain an intermediate matrix;

[0022] Subtract every other row of the intermediate matrix to obtain the gradient matrix, and only retain the elements greater than 0 in the gradient matrix to obtain the positive gradient matrix;

[0023] Accumulate all values ​​of each row in the positive gradient matrix, find the maximum value of all rows from the column vectors obtained by the accumulation as the grayscale change value of the circle center value, and store the grayscale change value and the corresponding row number in the corresponding positions of the maximum value matrix and the radius matrix respectively;

[0024] Traverse the circle center search range, take the value of each circle center in the circle center search range, repeat the above steps respectively, and obtain the maximum value matrix and radius matrix;

[0025] The center of the circle corresponding to the maximum grayscale change value in the maximum value matrix is ​​taken as the center of the image, and the number of rows in the radius matrix corresponding to the maximum grayscale change value is taken as the radius of the sub-image.

[0026] Furthermore, the iris image is divided into first to fourth sub-images according to the central angles of [0,π / 2], [π / 2,π], [π,3π / 2], and [3π / 2,2π]. When the annular region image is expanded into a rectangular region image, the central angles of the annular regions of the first to fourth sub-images are [-π / 5,π / 2+π / 5], [π / 2-π / 5,π+π / 5], [π-π / 5,3π / 2+π / 5], and [3π / 2-π / 5,π+π / 5], respectively.

[0027] The center search range is [x0-10, x0+10], [y0-10, y0+10], and the radius search range is [r1, r2], where (x0, y0) is the initial positioning center, r1=r0-25, r2=r0+25, and r0 is the initial positioning radius.

[0028] Furthermore, the method further comprises:

[0029] The horizontal coordinate variance and the vertical coordinate variance of the center of each sub-image are calculated. The horizontal coordinate variance and the vertical coordinate variance represent the degree of left-right pupil strabismus and the degree of up-down pupil strabismus, respectively. If the horizontal coordinate variance is greater than a set first threshold or the vertical coordinate variance is greater than a set second threshold, the iris image is determined to be unqualified.

[0030] Furthermore, the iris image is divided into a plurality of sub-images according to a certain central angle of the circle with the initial positioning center as the center, and the above also includes:

[0031] Binarizing the iris image using a set binarization threshold to obtain a binarized image;

[0032] Performing a dilation operation on the binary image to locate the light spot;

[0033] The pixels in the light spot on the iris image are interpolated biquadratically with the adjacent pixels.

[0034] In a second aspect, the present invention provides a strabismus pupil locating device, the device comprising:

[0035] An initial positioning module is used to perform initial positioning on the iris image by taking the pupil boundary as a circle and obtaining an initial positioning circle center and an initial positioning radius;

[0036] A sub-image division module is used to divide the iris image into a plurality of sub-images according to a certain central angle of the circle with the initial positioning center as the center;

[0037] A search range setting module is used to set the center search range and radius search range based on the initial positioning center and initial positioning radius;

[0038] A traversal module is used to traverse the center search range and radius search range for each sub-image, and take the center value and radius value corresponding to the maximum grayscale change as the center and radius of the sub-image;

[0039] The pupil boundary acquisition module is used to obtain the arc segment of each sub-image according to the center, radius and center angle of each sub-image, and splice the arc segments of all sub-images together to obtain the pupil boundary.

[0040] Furthermore, the traversal module includes:

[0041] An image expansion unit is used to expand, for each sub-image, a circular area image corresponding to the sub-image with a center value (x, y) and a radius search range of [r1, r2] into a rectangular area image;

[0042] A convolution unit is used to construct a filter of size n*1, and perform a convolution operation on the filter and the rectangular area image to obtain an intermediate matrix;

[0043] The gradient calculation unit is used to subtract every other row of the intermediate matrix to obtain a gradient matrix, and only retain the elements greater than 0 in the gradient matrix to obtain a positive gradient matrix;

[0044] an accumulation unit, configured to accumulate all values ​​of each row in the positive gradient matrix, find the maximum value of all rows from a column vector obtained by the accumulation as the grayscale change value of the circle center value, and store the grayscale change value and the corresponding row number in corresponding positions of the maximum value matrix and the radius matrix respectively;

[0045] The first traversal unit is used to traverse the circle center search range, take the value of each circle center within the circle center search range, and repeat the above image expansion unit, convolution unit, gradient calculation unit and accumulation unit to obtain the maximum value matrix and radius matrix;

[0046] The center and radius determination unit is used to take the center value corresponding to the maximum grayscale change value in the maximum value matrix as the center of the image, and the row number in the radius matrix corresponding to the maximum grayscale change value as the radius of the sub-image.

[0047] Furthermore, the iris image is divided into first to fourth sub-images according to the central angles of [0,π / 2], [π / 2,π], [π,3π / 2], and [3π / 2,2π]. When the annular region image is expanded into a rectangular region image, the central angles of the annular regions of the first to fourth sub-images are [-π / 5,π / 2+π / 5], [π / 2-π / 5,π+π / 5], [π-π / 5,3π / 2+π / 5], and [3π / 2-π / 5,π+π / 5], respectively.

[0048] The center search range is [x0-10, x0+10], [y0-10, y0+10], and the radius search range is [r1, r2], where (x0, y0) is the initial positioning center, r1=r0-25, r2=r0+25, and r0 is the initial positioning radius.

[0049] Furthermore, the device further comprises:

[0050] The strabismus degree determination module is used to calculate the horizontal coordinate variance and the vertical coordinate variance of the center of each sub-image. The horizontal coordinate variance and the vertical coordinate variance respectively represent the degree of left-right strabismus and up-down strabismus of the pupil. If the horizontal coordinate variance is greater than a set first threshold or the vertical coordinate variance is greater than a set second threshold, the iris image is determined to be unqualified.

[0051] Furthermore, the device further comprises:

[0052] A binarization module, configured to perform binarization processing on the iris image using a set binarization threshold to obtain a binarized image;

[0053] An image expansion module, configured to perform an expansion operation on the binary image to locate a light spot;

[0054] The interpolation module is used to perform biquadratic interpolation between the pixel points located within the light spot on the iris image and the adjacent pixel points.

[0055] In a third aspect, the present invention provides a computer-readable storage medium for strabismus pupil localization, comprising a memory for storing processor-executable instructions, wherein the instructions, when executed by the processor, implement the steps of the strabismus pupil localization method described in the first aspect.

[0056] In a fourth aspect, the present invention provides a device for strabismus pupil locating, comprising at least one processor and a memory storing computer-executable instructions, wherein when the processor executes the instructions, the steps of the strabismus pupil locating method described in the first aspect are implemented.

[0057] The present invention has the following beneficial effects:

[0058] This invention divides the pupil arc of strabismus into several angular intervals and performs positioning within each angular interval, improving the accuracy of strabismus pupil positioning and ultimately improving the accuracy of iris recognition. Furthermore, when positioning within each angular interval, there is no need to re-position the center and radius of the circle. Instead, the initial positioning center and radius are used as a reference for traversal search around the area, narrowing the positioning range and expediting positioning speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a schematic diagram of an iris image of a normal iris;

[0060] Figure 2 Schematic diagram of an iris image with a strabismus pupil;

[0061] Figure 3 Flowchart of one embodiment of the strabismus pupil locating method of the present invention;

[0062] Figure 4 This is a schematic diagram of the pupil after initial positioning;

[0063] Figure 5 Schematic diagram of the pupil boundary located on the first sub-image;

[0064] Figure 6 A schematic diagram of the pupil boundary located by the strabismus pupil location method of the present invention;

[0065] Figure 7 Schematic diagram of the iris image after removing the light spots;

[0066] Figure 8 is a specific flow chart of S400 in the strabismus pupil locating method of the present invention;

[0067] Figure 9 is a flow chart of another embodiment of the strabismus pupil locating method of the present invention;

[0068] Figure 10 A flow chart of a method for removing light spots;

[0069] Figure 11 is a schematic diagram of an embodiment of a strabismus pupil locating device of the present invention;

[0070] Figure 12 Schematic diagram of a traversal module in the strabismus pupil locating device of the present invention;

[0071] Figure 13 is a schematic diagram of another embodiment of the strabismus pupil locating device of the present invention;

[0072] Figure 14 Schematic diagram of the module for removing light spots. DETAILED DESCRIPTION

[0073] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0074] Example 1:

[0075] The embodiment of the present invention provides a strabismus pupil positioning method for positioning the strabismus pupil in iris recognition. Of course, the method can also be used to position the normal non-strabismus pupil. Figure 3 As shown, the method includes:

[0076] S100: Initially locating the pupil boundary as a circle on the iris image to obtain an initial positioning circle center and an initial positioning radius.

[0077] The present invention does not limit the method of initial positioning. In one example, the Sobel algorithm is first used to perform gradient processing and sharpening on the image to highlight the pupil boundary. Then, the boundary information is used to use the curve fitting method of symmetric radial transformation to realize initial positioning by treating the pupil boundary as a circle.

[0078] The specific effects of initial positioning are as follows: Figure 4 As shown, the white arc line is the initially located pupil boundary, the white cross is the initially located center point, the pupil boundary is the circular boundary, the center coordinates are (x0, y0), and the radius is r0.

[0079] The methods for gradient processing and sharpening of images can also include the Canny operator, and the curve fitting method can also include the least squares method, Hough transform, etc.

[0080] S200: Dividing the iris image into a plurality of sub-images according to a certain central angle of the circle, with the initially positioned center of the circle as the center.

[0081] Strabismus can cause the pupil boundary to distort, rendering it non-circular and inaccurate. To address this issue, the present invention divides the iris image into several sub-images, segmenting the pupil boundary into multiple segments. Because the pupil boundary can be viewed as an irregular arc, each end of the arc boundary is treated as a circle, ensuring circle fitting accuracy. These multiple segments are then spliced ​​together to form the entire pupil boundary.

[0082] S300: Based on the initial positioning center and initial positioning radius, a center search range and a radius search range are set.

[0083] The coordinates of the initial positioning center are (x0, y0), and the initial positioning radius is r0. Based on the initial positioning center and the initial positioning radius, a certain range is expanded around to obtain the center search range and radius search range.

[0084] For example, the search range for the center is [x0-10, x0+10], [y0-10, y0+10], i.e., the center coordinates in the x-direction range are [x0-10, x0+10], and in the y-direction range are [y0-10, y0+10]. The search range for the radius is [r1, r2], where r1 = r0-25 and r2 = r0+25.

[0085] S400: For each sub-image, traverse the circle center search range and the radius search range, and use the circle center value and radius value corresponding to the maximum grayscale change as the circle center and radius of the sub-image.

[0086] In practice, you can first fix the center value, traverse the radius search range, and find the maximum grayscale change of the center value within the radius search range as the grayscale change value of the center, and obtain the radius value corresponding to this grayscale change value. Then, traverse the center search range and repeat the above operation for each center value within the center search range. Finally, the center value and radius value corresponding to the maximum grayscale change value are used as the center and radius of the sub-image.

[0087] For example, first fix the center of the circle and traverse the radius from r1 to r2. Find the maximum grayscale change within the radius traversal range as the grayscale change value of the center of the circle and its corresponding radius value. Then change the center of the circle again and traverse the search range of the center of the circle [x0-10, x0+10], [y0-10, y0+10]. Each center value is assigned a grayscale change value and radius value. Find the global maximum grayscale change value and its corresponding center value and radius value, which are the center and radius of a sub-image.

[0088] Repeat the above steps for each sub-image to obtain the center and radius of each sub-image.

[0089] When traversing the center search range and the radius search range, the traversal step can be set. Generally, the traversal step is 1. In situations where the accuracy requirement is not high, the traversal step can be an integer greater than 1.

[0090] S500: Obtaining an arc segment of each sub-image according to the center, radius, and center angle of each sub-image, and splicing the arc segments of all sub-images together to obtain a pupil boundary.

[0091] Each sub-image is given a center and radius, and according to the center angle of the sub-image, an arc can be obtained, such as Figure 5 As shown. The arcs of multiple sub-images are stitched together to obtain the pupil boundary, as shown Figure 6 shown.

[0092] The strabismus pupil location method of the present invention divides the deformed pupil arc into several angular intervals and performs location determination within each angular interval, thereby improving the accuracy of strabismus pupil location and ultimately the accuracy of iris recognition. Furthermore, when locating within each angular interval, the center and radius of the circle do not need to be re-located. Instead, the initial location is used as a reference for traversing the surrounding area, narrowing the location range and expediting the location determination process.

[0093] The present invention does not limit the method of dividing the sub-images. In one example, the iris image is divided into first to fourth sub-images according to the central angles of [0,π / 2], [π / 2,π], [π,3π / 2], and [3π / 2,2π]. The first to fourth sub-images are respectively in the first to fourth quadrants.

[0094] Based on the above sub-image division method, such as Figure 8 As shown, the aforementioned S400 includes:

[0095] S410: For each sub-image, expand the annular region image corresponding to the sub-image with the center value (x, y) and the radius search range being [r1, r2] into a rectangular region image.

[0096] Taking the first sub-image as an example, the annular region image with the center value (x, y), the radius search range [r1, r2], and the center angle [0, π / 2] is expanded into a rectangular region image.

[0097] The image is an annular region with a radius of [r1, r2] and a center angle of [0, π / 2]. The annular region image is expanded into a row (r2-r1+1), which is divided into several rectangular region images. The number of rows in the rectangular region image varies depending on the image resolution. For example, the number of rows can be 60.

[0098] Furthermore, the center angle of the first sub-image is [0, π / 2]. When expanded into a rectangle, the center angle of the annular region can be expanded outward by π / 5 on both ends of [0, π / 2], so that the center angle of the annular region is [-π / 5, π / 2+π / 5]. This utilizes some information outside the edge of the first sub-image, making positioning more accurate. Similarly, the center angles of the annular regions corresponding to the second to fourth sub-images are [π / 2-π / 5, π+π / 5], [π-π / 5, 3π / 2+π / 5], and [3π / 2-π / 5, π+π / 5], respectively.

[0099] The above image can be expanded using a normalization method. The specific normalization method is:

[0100] Convert the Cartesian coordinates (x, y) of the image to polar coordinates (r, θ) using the following formula:

[0101]

[0102] Among them, r∈[0,1], θ∈[0,2π], (x in ,y in ), (x out ,y out ) represent the starting point and end point in the θ direction respectively.

[0103] S420: Construct a filter of size n*1, and perform a convolution operation on the filter and the rectangular region image to obtain an intermediate matrix.

[0104] The filter is a matrix with n rows and 1 column. The convolution operation is performed on the rectangular area image through the filter, and the grayscale value of each row of the rectangular area image is replaced by the statistical value of several rows around the row. The value of n can be set as needed.

[0105] For example, the filter is [1,1,1,1,1]', and the grayscale value of each row after convolution is the statistical value of the surrounding five rows. After convolution, the present invention can also perform a point division operation on the convolution result, for example, the grayscale value of each point is divided by 5, so that the grayscale value of each point is the average of the grayscale values ​​of the five valid rows nearby.

[0106] S430: Subtract every other row of the intermediate matrix to obtain a gradient matrix, and retain only the elements greater than 0 in the gradient matrix to obtain a positive gradient matrix.

[0107] Interleaved subtraction means subtracting the first row from the third row, the second row from the fourth row, and so on, to obtain the gradient matrix. Since the pupil grayscale is smaller than the iris grayscale, only values ​​> 0 in the gradient matrix are taken to obtain the positive gradient matrix.

[0108] S440: Accumulate all the values ​​of each row in the positive gradient matrix, find the maximum value of all rows from the column vectors obtained by accumulation as the grayscale change value of the center value of the circle, and store the grayscale change value and the corresponding row number in the corresponding positions of the maximum value matrix and the radius matrix respectively.

[0109] In this step, each row of the positive gradient matrix is ​​accumulated, and then the values ​​of all rows of the accumulated column vector are compared. The maximum value of all rows is the grayscale change value, and the row number corresponding to the maximum value of all rows is the corresponding position of the pupil boundary, that is, the radius.

[0110] S450: Traverse the circle center search range, take the value of each circle center within the circle center search range, repeat the above steps S410 to S440 respectively, and obtain the maximum value matrix and radius matrix. Each position of the maximum value matrix and radius matrix corresponds to a circle center value.

[0111] S460: taking the center value corresponding to the maximum grayscale change value in the maximum value matrix as the center value of the image, and the row number in the radius matrix corresponding to the maximum grayscale change value as the radius of the sub-image.

[0112] For the first sub-image, the maximum value in the maximum value matrix is ​​the maximum value of the global grayscale change, and the row number in the corresponding radius matrix is ​​the corresponding position of the pupil boundary, that is, the radius. The center value corresponding to the maximum value of the global grayscale change is the center of the circle, such as Figure 5 As shown. The center and radius of the first to fourth sub-images are (x1, y1, r1), (x2, y2, r2), (x3, y3, r3), (x4, y4, r4), respectively. Figure 6 shown.

[0113] As an improvement to the embodiment of the present invention, Figure 9 As shown, the method further includes:

[0114] S600: Calculate the horizontal coordinate variance and the vertical coordinate variance of the center of all sub-images. The horizontal coordinate variance and the vertical coordinate variance represent the degree of left-right pupil strabismus and the degree of up-down pupil strabismus, respectively. If the horizontal coordinate variance is greater than a set first threshold or the vertical coordinate variance is greater than a set second threshold, it is determined that the iris image does not meet the requirements.

[0115] This step uses the center coordinates of multiple sub-images to determine whether the iris image has strabismus and the degree of strabismus. Taking the aforementioned (x1, y1, r1), (x2, y2, r2), (x3, y3, r3), (x4, y4, r4) as an example, the specific judgment criteria are as follows:

[0116] If the variance of the four numbers (y1, y2, y3, y4) is greater than the second threshold M, it is considered that the upper and lower strabismus is too large and does not meet the requirements; if the variance of the four numbers (x1, x2, x3, x4) is greater than the first threshold N, it is considered that the left and right strabismus is too large and does not meet the requirements.

[0117] Among them, M and N can be set according to actual requirements. Generally, the thresholds M and N are set according to the situation when iris recognition can be successfully matched. When the horizontal coordinate variance or the vertical coordinate variance is greater than N and M respectively, the strabismus is too large and iris recognition cannot be successfully matched.

[0118] There are often light spots in iris images. Because the gradient of light spots on the pupil boundary is relatively large, it will affect the pupil positioning effect. In order to remove the influence of light spot noise on boundary fitting, Figure 10 As shown, the present invention further includes, after the initial positioning and before the sub-image division:

[0119] S110: performing binarization processing on the iris image using a set binarization threshold to obtain a binarized image.

[0120] For example, the binarization threshold is set to 250, and binarization is performed with 250 as the threshold. Values ​​greater than 250 are set to 1 (or 255), and the others are set to 0 to obtain a binarized image. The position of 1 on the binarized image represents the light spot, which can roughly determine the position and size of the light spot.

[0121] S120: Performing a dilation operation on the binary image to locate the light spot.

[0122] Dilation is similar to "domain expansion," expanding the highlighted or white areas of the image. The resulting image is larger than the original image's highlighted areas. This step removes holes within the spot or jagged edges.

[0123] S130: Perform biquadratic interpolation on the pixel points within the light spot on the iris image and the adjacent pixel points.

[0124] This step uses the pixel values ​​around the light spot to perform calculations instead of the values ​​within the light spot to achieve the purpose of removing the light spot. The effect after removing the light spot is as follows: Figure 7 shown.

[0125] This invention segments the pupil's deformed arc into multiple angle intervals and positions each pupil boundary as an arc. This improves the accuracy of pupil location in strabismus and ultimately improves iris recognition accuracy. Furthermore, for iris images with severe strabismus, where severe iris deformation significantly impacts recognition, the degree of strabismus in the image can be determined based on multiple sets of pupil location coordinates.

[0126] Example 2:

[0127] The embodiment of the present invention provides a strabismus pupil positioning device, such as Figure 11 As shown, the device includes:

[0128] The initial positioning module 100 is used to perform initial positioning on the iris image by taking the pupil boundary as a circle to obtain an initial positioning circle center and an initial positioning radius.

[0129] The sub-image division module 200 is used to divide the iris image into a plurality of sub-images according to a certain central angle of the circle, with the initially positioned center of the circle as the center.

[0130] The search range setting module 300 is used to set a circle center search range and a radius search range based on the initial positioning circle center and the initial positioning radius.

[0131] The traversal module 400 is used to traverse the center search range and the radius search range for each sub-image, and use the center value and radius value corresponding to the maximum grayscale change as the center and radius of the sub-image.

[0132] The pupil boundary acquisition module 500 is used to obtain the arc segment of each sub-image according to the center, radius and center angle of each sub-image, and to splice the arc segments of all sub-images together to obtain the pupil boundary.

[0133] This invention divides the pupil arc of strabismus into several angular intervals and performs positioning within each angular interval, improving the accuracy of strabismus pupil positioning and ultimately improving the accuracy of iris recognition. Furthermore, when positioning within each angular interval, there is no need to re-position the center and radius of the circle. Instead, the initial positioning center and radius are used as a reference for traversal search around the area, narrowing the positioning range and expediting positioning speed.

[0134] The present invention does not limit the method of dividing the sub-images. In one example, the iris image is divided into first to fourth sub-images according to the central angles of [0,π / 2], [π / 2,π], [π,3π / 2], and [3π / 2,2π]. The first to fourth sub-images are respectively in the first to fourth quadrants.

[0135] Based on the above sub-image division method, such as Figure 12 As shown, the aforementioned traversal module 400 includes:

[0136] The image expansion unit 410 is configured to expand, for each sub-image, a circular region image corresponding to the sub-image with a center value (x, y) and a radius search range of [r1, r2] into a rectangular region image.

[0137] The convolution unit 420 is used to construct a filter of size n*1, and perform a convolution operation on the filter and the rectangular area image to obtain an intermediate matrix.

[0138] The gradient calculation unit 430 is used to subtract every other row of the intermediate matrix to obtain a gradient matrix, and retain only the elements greater than 0 in the gradient matrix to obtain a positive gradient matrix.

[0139] The accumulation unit 440 is used to accumulate all the values ​​of each row in the positive gradient matrix, find the maximum value of all rows from the column vectors obtained by the accumulation as the grayscale change value of the center value of the circle, and store the grayscale change value and the corresponding row number in the corresponding positions of the maximum value matrix and the radius matrix respectively.

[0140] The first traversal unit 450 is used to traverse the center search range, and for each center value within the center search range, repeat the above-mentioned image expansion unit, convolution unit, gradient calculation unit and accumulation unit to obtain the maximum value matrix and radius matrix. Each position of the maximum value matrix and radius matrix corresponds to a center value.

[0141] The center and radius determination unit 460 is used to take the center value corresponding to the maximum grayscale change value in the maximum value matrix as the center of the image, and the row number in the radius matrix corresponding to the maximum grayscale change value as the radius of the sub-image.

[0142] In one example, when the annular area image is expanded into a rectangular area image, the central angles of the annular areas of the first to fourth sub-images are [-π / 5,π / 2+π / 5], [π / 2-π / 5,π+π / 5], [π-π / 5,3π / 2+π / 5], and [3π / 2-π / 5,π+π / 5], respectively.

[0143] The center search range is [x0-10, x0+10], [y0-10, y0+10], and the radius search range is [r1, r2], where (x0, y0) is the initial positioning center, r1 = r0-25, r2 = r0+25, and r0 is the initial positioning radius.

[0144] As an improvement to the embodiment of the present invention, Figure 13 As shown, the device also includes:

[0145] The strabismus degree determination module 600 is configured to calculate the horizontal and vertical variances of the centers of all sub-images. The horizontal and vertical variances represent the degree of left-right and up-down pupil strabismus, respectively. If the horizontal variance is greater than a first threshold or the vertical variance is greater than a second threshold, the iris image is determined to be unqualified.

[0146] To remove the effect of light spots, Figure 14 As shown, the device of the present invention also includes:

[0147] The binarization module 110 is configured to perform binarization processing on the iris image using a set binarization threshold to obtain a binarized image.

[0148] The image expansion module 120 is used to perform an expansion operation on the binary image to locate the light spot.

[0149] The interpolation module 130 is configured to perform biquadratic interpolation between the pixel points within the light spot and the adjacent pixel points on the iris image.

[0150] This invention segments the pupil's deformed arc into multiple angle intervals and positions each pupil boundary as an arc. This improves the accuracy of pupil location in strabismus and ultimately improves iris recognition accuracy. Furthermore, for iris images with severe strabismus, where severe iris deformation significantly impacts recognition, the degree of strabismus in the image can be determined based on multiple sets of pupil location coordinates.

[0151] The device provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment 1. For the sake of brevity, any matters not mentioned in the device embodiment can be referred to the corresponding contents in the aforementioned method embodiment 1. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the devices and units described above can all refer to the corresponding processes in the aforementioned method embodiment 1 and will not be repeated here.

[0152] Example 3:

[0153] The method described in Example 1 above, provided by the present invention, can implement the business logic through a computer program and record it on a storage medium. The storage medium can be read and executed by a computer to achieve the effects of the solution described in Example 1 of this specification. Therefore, the present invention also provides a computer-readable storage medium for strabismus pupil localization, including a memory for storing processor-executable instructions. When the instructions are executed by the processor, the steps of the strabismus pupil localization method described in Example 1 are implemented.

[0154] This invention divides the pupil arc of strabismus into several angular intervals and performs positioning within each angular interval, improving the accuracy of strabismus pupil positioning and ultimately improving the accuracy of iris recognition. Furthermore, when positioning within each angular interval, there is no need to re-position the center and radius of the circle. Instead, the initial positioning center and radius are used as a reference for traversal search around the area, narrowing the positioning range and expediting positioning speed.

[0155] The storage medium may include a physical device for storing information, typically digitizing the information and then storing it in a medium utilizing electrical, magnetic, or optical means. Examples of such storage media include: devices that use electrical energy to store information, such as various types of memory, such as RAM and ROM; devices that use magnetic energy to store information, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memories, and USB flash drives; and devices that use optical means to store information, such as CDs or DVDs. Of course, there are also other types of readable storage media, such as quantum memories and graphene memories.

[0156] The storage medium described above may also include other implementation methods according to the description of method embodiment 1. The implementation principle and technical effects produced by this embodiment are the same as those of the aforementioned method embodiment 1. For details, please refer to the description of the relevant method embodiment 1 and will not be repeated here.

[0157] Example 4:

[0158] The present invention also provides a device for strabismus pupil localization. The device may be a standalone computer or may include an actual operating device that utilizes one or more methods or one or more embodiments of the present specification. The device may include at least one processor and a memory storing computer-executable instructions. When the processor executes the instructions, the steps of the strabismus pupil localization method described in any one or more of Embodiment 1 are implemented.

[0159] This invention divides the pupil arc of strabismus into several angular intervals and performs positioning within each angular interval, improving the accuracy of strabismus pupil positioning and ultimately improving the accuracy of iris recognition. Furthermore, when positioning within each angular interval, there is no need to re-position the center and radius of the circle. Instead, the initial positioning center and radius are used as a reference for traversal search around the area, narrowing the positioning range and expediting positioning speed.

[0160] The above-mentioned device can also include other implementation methods according to the description of method embodiment 1. The implementation principle and technical effects produced by this embodiment are the same as those of the aforementioned method embodiment 1. For details, please refer to the description of the relevant method embodiment 1 and will not be repeated here.

[0161] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, intended to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can, within the technical scope disclosed by the present invention, modify or readily conceive of variations to the technical solutions described in the above-described embodiments, or substitute equivalently for some of the technical features thereof. Such modifications, variations, or substitutions do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. They should all be encompassed within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A strabismus pupil location method, characterized in that: The method comprises: The pupil boundary is initially positioned as a circle on the iris image to obtain the initial positioning circle center and initial positioning radius; Taking the initially positioned circle center as the center, the iris image is divided into a plurality of sub-images according to a certain central angle; Based on the initial positioning center and initial positioning radius, set the center search range and radius search range; For each sub-image, traverse the center search range and radius search range, and take the center value and radius value corresponding to the maximum grayscale change as the center and radius of the sub-image; Obtain the arc segment of each sub-image according to the center, radius and central angle of each sub-image, and splice the arc segments of all sub-images together to obtain the pupil boundary; For each sub-image, traversing the circle center search range and the radius search range, and taking the circle center value and radius value corresponding to the maximum grayscale change as the circle center and radius of the sub-image, including: For each sub-image, the center of the circle corresponding to the sub-image is (x, y) and the radius search range is [r 1, r2] is expanded into a rectangular area image; Constructing a filter of size n*1, and performing a convolution operation on the filter and the rectangular area image to obtain an intermediate matrix; Subtract every other row of the intermediate matrix to obtain the gradient matrix, and only retain the elements greater than 0 in the gradient matrix to obtain the positive gradient matrix; Accumulate all values ​​of each row in the positive gradient matrix, find the maximum value of all rows from the column vectors obtained by the accumulation as the grayscale change value of the circle center value, and store the grayscale change value and the corresponding row number in the corresponding positions of the maximum value matrix and the radius matrix respectively; Traverse the circle center search range, take the value of each circle center in the circle center search range, repeat the above steps respectively, and obtain the maximum value matrix and radius matrix; The center search range and the radius search range are obtained by expanding the set range around the initial positioning center and the initial positioning radius respectively; The center of the circle corresponding to the maximum grayscale change value in the maximum value matrix is ​​taken as the center of the image, and the number of rows in the radius matrix corresponding to the maximum grayscale change value is taken as the radius of the sub-image.

2. The strabismus pupil locating method according to claim 1, wherein: The iris image is divided into first to fourth sub-images according to the central angles of [0,π / 2], [π / 2,π], [π,3π / 2], and [3π / 2,2π]. When the annular region image is expanded into a rectangular region image, the central angles of the annular regions of the first to fourth sub-images are [-π / 5,π / 2+π / 5], [π / 2-π / 5,π+π / 5], [π-π / 5,3π / 2+π / 5], and [3π / 2-π / 5,π+π / 5], respectively. The center search range is [x0-10, x0+10], [y0-10, y0+10], and the radius search range is [r 1, r2], where (x0, y0) is the initial positioning center, r1 = r0-25, r2 = r0+25, and r0 is the initial positioning radius.

3. The strabismus pupil locating method according to claim 1, wherein: The method further comprises: The horizontal coordinate variance and the vertical coordinate variance of the center of each sub-image are calculated. The horizontal coordinate variance and the vertical coordinate variance represent the degree of left-right pupil strabismus and the degree of up-down pupil strabismus, respectively. If the horizontal coordinate variance is greater than a set first threshold or the vertical coordinate variance is greater than a set second threshold, the iris image is determined to be unqualified.

4. The strabismus pupil locating method according to any one of claims 1 to 3, characterized in that: The method of dividing the iris image into a plurality of sub-images based on a certain central angle of the circle with the initially positioned center as the center also includes: Binarizing the iris image using a set binarization threshold to obtain a binarized image; Performing a dilation operation on the binary image to locate the light spot; The pixels in the light spot on the iris image are interpolated biquadratically with the adjacent pixels.

5. A strabismus pupil locating device, characterized in that: The device comprises: An initial positioning module is used to perform initial positioning on the iris image by taking the pupil boundary as a circle and obtaining an initial positioning circle center and an initial positioning radius; A sub-image division module is used to divide the iris image into a plurality of sub-images according to a certain central angle of the circle with the initial positioning center as the center; A search range setting module is used to set the center search range and radius search range based on the initial positioning center and initial positioning radius; A traversal module is used to traverse the center search range and radius search range for each sub-image, and take the center value and radius value corresponding to the maximum grayscale change as the center and radius of the sub-image; A pupil boundary acquisition module is used to obtain the arc segment of each sub-image according to the center, radius and central angle of each sub-image, and to splice the arc segments of all sub-images together to obtain the pupil boundary; The traversal module includes: The image expansion unit is used to set the center of the circle corresponding to each sub-image to (x, y) and the radius search range to [r 1, r2] is expanded into a rectangular area image; A convolution unit is used to construct a filter of size n*1, and perform a convolution operation on the filter and the rectangular area image to obtain an intermediate matrix; The gradient calculation unit is used to subtract every other row of the intermediate matrix to obtain a gradient matrix, and only retain the elements greater than 0 in the gradient matrix to obtain a positive gradient matrix; an accumulation unit, configured to accumulate all values ​​of each row in the positive gradient matrix, find the maximum value of all rows from a column vector obtained by the accumulation as the grayscale change value of the circle center value, and store the grayscale change value and the corresponding row number in corresponding positions of the maximum value matrix and the radius matrix respectively; The first traversal unit is used to traverse the circle center search range, take the value of each circle center within the circle center search range, and repeat the above image expansion unit, convolution unit, gradient calculation unit and accumulation unit to obtain the maximum value matrix and radius matrix; The center search range and the radius search range are obtained by expanding the set range around the initial positioning center and the initial positioning radius respectively; The center and radius determination unit is used to take the center value corresponding to the maximum grayscale change value in the maximum value matrix as the center of the image, and the row number in the radius matrix corresponding to the maximum grayscale change value as the radius of the sub-image.

6. The strabismus pupil locating device according to claim 5, characterized in that: The device further comprises: The strabismus degree determination module is used to calculate the horizontal coordinate variance and the vertical coordinate variance of the center of each sub-image. The horizontal coordinate variance and the vertical coordinate variance respectively represent the degree of left-right strabismus and up-down strabismus of the pupil. If the horizontal coordinate variance is greater than a set first threshold or the vertical coordinate variance is greater than a set second threshold, the iris image is determined to be unqualified.

7. A computer-readable storage medium for strabismus pupil localization, characterized in that: The invention comprises a memory for storing processor-executable instructions, wherein when the instructions are executed by the processor, the steps of the strabismus pupil locating method according to any one of claims 1 to 4 are implemented.

8. A device for locating the pupil of strabismus, characterized in that: The method comprises at least one processor and a memory storing computer-executable instructions, wherein when the processor executes the instructions, the steps of the strabismus pupil locating method according to any one of claims 1 to 4 are implemented.

Images