Palm biological feature fusion recognition method and device, storage medium and equipment

Abstract

The invention discloses a palm biological feature fusion recognition method and device, a storage medium and equipment, and belongs to the field of biological feature recognition. According to the method, the characteristic that the distribution positions of the palm print main line and the palm vein blood vessels are relatively fixed is utilized, the intersection points of the palm print main line and the palm vein blood vessels serve as the feature points, the feature point set used for registration and recognition is constructed, feature-level fusion of the palm prints and the palm veins is achieved, and comparison is conducted through the fused feature point set, so that the registration accuracy is improved. Therefore, palm biological feature fusion recognition is realized. According to the method, the advantages that the palmprint principal line for palmprint recognition is thicker, obvious and stable in feature and not prone to interference compared with other features, the blood vessel near-infrared imaging for palmprint vein recognition is hidden and not prone to interference, and the advantages of correlation and dependence of relative position distribution of palmprint and palmprint vein are combined, so that the accuracy of palmprint and palmprint vein recognition is improved, and the accuracy of palmprint and palmprint vein recognition is improved. The recognition precision and the recognition reliability of biological characteristics are greatly improved, and accurate recognition of palm prints and palm veins is achieved.

Description

Technical Field

[0001] This application relates to the field of biometric recognition, and in particular to a method, apparatus, storage medium and device for palm biometric recognition fusion. Background Technology

[0002] In today's information age, accurately identifying an individual and protecting information security has become a critical social issue that must be addressed. While single-modal biometric technologies have achieved good recognition results, they still have many shortcomings in terms of applicable scenarios, privacy sensitivity, and recognition accuracy. Therefore, researching multimodal biometric recognition technologies, and further improving the accuracy, robustness, and security of biometric recognition technologies by selecting and combining biometric modalities and fusion methods, while expanding the applicable scenarios, has become a key focus.

[0003] Non-contact palmprint and palm vein fusion recognition technology, as a new generation of multimodal biometric recognition technology, combines the identification advantages of palmprint recognition and the liveness detection advantages of palm vein recognition. Compared with single-feature recognition technology, it has higher recognition accuracy and security. Furthermore, the non-contact, one-time simultaneous acquisition of palmprint and palm vein image information also brings users a convenient, fast, and safe user experience.

[0004] However, existing palm print and palm vein fusion recognition technologies are basically simple superposition and fusion methods, which do not significantly improve the accuracy and reliability of biometric recognition. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application provides a method, apparatus, storage medium, and device for palm biometric fusion recognition, which greatly improves the recognition accuracy and reliability of biometrics and enables accurate recognition of palm prints and palm veins.

[0006] The technical solution provided in this application is as follows:

[0007] In a first aspect, this application provides a method for palm biometric fusion recognition, the method comprising:

[0008] Simultaneously acquire palm print and palm vein images of the same hand of the same user;

[0009] Based on the palm print image and the palm vein image, obtain the effective area of ​​the palm print and the effective area of ​​the palm vein.

[0010] The effective area of ​​the palm print and the effective area of ​​the palm vein are located at the same position on the user's palm.

[0011] The main palm print line is extracted based on the effective area of ​​the palm print.

[0012] The palm vein vessels were extracted based on the effective area of ​​the palm vein.

[0013] In the same coordinate system, the intersection points of the palm print main line and the palm vein are extracted to obtain the set of feature points to be identified;

[0014] The set of feature points to be identified is matched with a pre-registered set of template feature points to achieve palm biometric fusion recognition.

[0015] Furthermore, in the same coordinate system, the intersection points of the palm print main line and the palm vein are extracted to obtain the set of feature points to be identified, including:

[0016] In the same coordinate system, find the point on the palm print main line with the same coordinate as the palm vein, and obtain the intersection point as the feature point to be identified;

[0017] On the same palm print line, connect each feature point to be identified in sequence from one end of the palm print line to the other.

[0018] For any feature point to be identified, calculate the angle between the vector from the feature point to the next feature point and the positive x-axis, and use it as the direction of the feature point to be identified;

[0019] The set of feature points to be identified is formed by the position coordinates, orientation, and palm print main line identifier of each feature point to be identified; wherein, the palm print main line identifier represents the palm print main line to which the feature point to be identified belongs.

[0020] Furthermore, the step of matching the set of feature points to be identified with a pre-registered template feature point set to achieve palm biometric fusion recognition includes:

[0021] Based on the palmprint main line identifiers of the feature points to be identified contained in the set of feature points to be identified and the palmprint main line identifiers of the template feature points contained in the set of template feature points, the feature points to be identified and the template feature points belonging to the same palmprint main line are determined.

[0022] The position coordinates and directions of the feature points to be identified and the template feature points belonging to the same palm print main line are matched respectively. The feature points to be identified and the template feature points that are matched in both position coordinates and direction are taken as the matching feature point pairs.

[0023] Calculate the percentage of the total number of to-be-identified feature points and template feature points included in the matching feature point pair to the total number of to-be-identified feature points included in the to-be-identified feature point set and the total number of template feature points included in the template feature point set;

[0024] Determine whether the percentage is greater than the set percentage threshold. If it is, the palm biometric fusion recognition passes; otherwise, the palm biometric fusion recognition fails.

[0025] Furthermore, the step of matching the position coordinates and directions of the feature points to be identified and the template feature points belonging to the same palm print main line, and taking the feature points to be identified and the template feature points that match in both position coordinates and direction as a matching feature point pair, includes:

[0026] The distance information between the feature point to be identified and the template feature point is calculated based on the position coordinates of the feature point to be identified and the template feature point belonging to the same palm print main line.

[0027] The distance information is compared with a preset distance range. If the distance information is within the distance range, the location coordinates are determined to be a successful match; otherwise, the location coordinates are determined to be a failed match.

[0028] Calculate the angle difference information between the feature points to be identified and the template feature points that belong to the same palm print main line;

[0029] The angle difference information is compared with a preset angle difference range. If the angle difference information is within the angle difference range, the direction matching is determined to be successful; otherwise, the direction matching is determined to be unsuccessful.

[0030] The feature points to be identified and the template feature points that pass both position coordinate matching and direction matching are used as matching feature point pairs.

[0031] Furthermore, the step of extracting the palm print main line based on the effective palm print area includes:

[0032] A contrast enhancement operation is performed on the effective area of ​​the palm print to obtain an enhanced image;

[0033] Palmprint feature points are extracted from the enhanced image to obtain a palmprint feature point map;

[0034] Based on the palmprint feature point map, generate the palmprint main line to obtain the palmprint main line map;

[0035] The palmprint main line image is refined to obtain a palmprint main line with a single pixel width.

[0036] Furthermore, the step of extracting the palm vein vessels based on the effective area of ​​the palm vein includes:

[0037] Median filtering is applied to the effective region of the palm vein;

[0038] Enhance the contrast of the image after median filtering;

[0039] The contrast-enhanced image is binarized according to the set binarization threshold.

[0040] The binarized image is thinned to obtain the palmar vein.

[0041] Furthermore, the method also includes:

[0042] Palm print recognition and palm vein recognition are performed based on the effective area of ​​the palm print and the effective area of ​​the palm vein, respectively.

[0043] Multimodal biometric recognition is performed based on the fusion recognition of palm biometric features, the palm print recognition, and the palm vein recognition.

[0044] Secondly, this application provides a palm biometric fusion recognition device, characterized in that the device comprises:

[0045] The image acquisition module is used to simultaneously acquire palm print images and palm vein images of the same hand of the same user;

[0046] The preprocessing module is used to obtain the effective area of ​​the palm print and the effective area of ​​the palm vein based on the palm print image and the palm vein image.

[0047] The effective area of ​​the palm print and the effective area of ​​the palm vein are located at the same position on the user's palm.

[0048] The palmprint main line extraction module is used to extract the palmprint main line based on the effective area of ​​the palmprint.

[0049] A palm vein extraction module is used to extract palm veins based on the effective area of ​​the palm vein.

[0050] The feature fusion module is used to extract the intersection points of the palm print main line and the palm vein vessels in the same coordinate system to obtain the set of feature points to be identified.

[0051] The identification module is used to match the set of feature points to be identified with a pre-registered set of template feature points to achieve palm biometric fusion identification.

[0052] Furthermore, the feature fusion module includes:

[0053] The feature point determination unit is used to find the point with the same coordinates on the palm print main line and the palm vein in the same coordinate system, and obtain the intersection point as the feature point to be identified.

[0054] The feature point connection unit is used to connect each feature point to be identified sequentially on the same palm print main line, in order from one end of the palm print main line to the other end.

[0055] The direction determination unit is used to calculate the angle between the vector from the feature point to the next feature point to be identified and the positive x-axis for any feature point to be identified, and use it as the direction of the feature point to be identified;

[0056] The feature point set determination unit is used to form the feature point set to be identified by taking the position coordinates, direction and palm print main line identifier of each feature point to be identified; wherein, the palm print main line identifier represents the palm print main line to which the feature point to be identified belongs.

[0057] Furthermore, the identification module includes:

[0058] The same main line feature point determination unit is used to determine the feature points to be identified and the template feature points belonging to the same palm print main line based on the palm print main line identifier of the feature points to be identified contained in the feature point set to be identified and the palm print main line identifier of the template feature points contained in the template feature point set.

[0059] The position and orientation matching unit is used to match the position coordinates and orientation of the feature points to be identified and the template feature points that belong to the same palm print main line, and to identify the feature points and template feature points that are matched in both position coordinates and orientation as a matching feature point pair;

[0060] The proportion calculation unit is used to calculate the proportion of the total number of the to-be-identified feature points and template feature points included in the matching feature point pair to the total number of the to-be-identified feature points included in the to-be-identified feature point set and the template feature points included in the template feature point set.

[0061] The judgment unit is used to determine whether the proportion is greater than the set proportion threshold. If it is, the palm biometric fusion recognition passes; otherwise, the palm biometric fusion recognition fails.

[0062] Furthermore, the position and orientation matching unit includes:

[0063] The distance calculation subunit is used to calculate the distance information between the feature point to be identified and the template feature point based on the position coordinates of the feature point to be identified and the template feature point belonging to the same palm print main line.

[0064] The location matching subunit is used to compare the distance information with a preset distance range. If the distance information is within the distance range, the location coordinates are determined to be matched successfully; otherwise, the location coordinates are determined to be matched unsuccessfully.

[0065] An angle difference calculation subunit is used to calculate the angle difference information between the feature point to be identified and the template feature point based on the direction of the feature point to be identified and the template feature point belonging to the same palm print main line.

[0066] The direction matching subunit is used to compare the angle difference information with a preset angle difference range. If the angle difference information is within the angle difference range, the direction matching is determined to be successful; otherwise, the direction matching is determined to be unsuccessful.

[0067] The matching feature point pair determination sub-unit is used to take the feature points to be identified and the template feature points that have passed both position coordinate matching and orientation matching as matching feature point pairs.

[0068] Furthermore, the palmprint main line extraction module includes:

[0069] A contrast enhancement unit is used to perform a contrast enhancement operation on the effective area of ​​the palm print to obtain an enhanced image;

[0070] A palmprint feature point extraction unit is used to extract palmprint feature points from the enhanced image to obtain a palmprint feature point map.

[0071] The palmprint main line extraction unit is used to generate palmprint main lines based on the palmprint characteristic point map, and obtain a palmprint main line map;

[0072] The palmprint main line refinement unit is used to refine the palmprint main line image to obtain a palmprint main line with a single pixel width.

[0073] Furthermore, the palm vein extraction module includes:

[0074] A filtering unit is used to perform median filtering on the effective area of ​​the palm vein;

[0075] A contrast enhancement unit is used to enhance the contrast of the image after median filtering.

[0076] The binarization unit is used to binarize the contrast-enhanced image according to the set binarization threshold.

[0077] The palm vein vessel refinement unit is used to refine the binarized image to obtain the palm vein vessels.

[0078] Furthermore, the device also includes:

[0079] A palm print and palm vein recognition module is used to perform palm print recognition and palm vein recognition based on the effective area of ​​the palm print and the effective area of ​​the palm vein, respectively.

[0080] A multimodal biometric recognition module is used to perform multimodal biometric recognition based on the fusion recognition of the palm biometric features, the palm print recognition, and the palm vein recognition.

[0081] Thirdly, this application provides a computer-readable storage medium for palm biometric fusion recognition, including a memory for storing processor-executable instructions, which, when executed by the processor, implement the steps of the palm biometric fusion recognition method described in the first aspect.

[0082] Fourthly, this application provides a device for palm biometric fusion recognition, characterized in that it includes at least one processor and a memory storing computer-executable instructions, wherein the processor executes the instructions to implement the steps of the palm biometric fusion recognition method described in the first aspect.

[0083] This application has the following beneficial effects:

[0084] This application leverages the relatively fixed distribution of palm print main lines and palm veins, using the intersections of these lines as feature points to construct a feature point set for registration and identification. This achieves feature-level fusion of palm print and palm vein characteristics, and compares the fused feature point set to realize palm biometric fusion identification. This application combines the advantages of palm print recognition (larger, more prominent, stable, and less susceptible to interference compared to other features) with palm vein recognition (concealed near-infrared imaging of blood vessels, less susceptible to interference) and the correlation and dependence between the relative positions of palm prints and palm veins, significantly improving the accuracy and reliability of biometric identification and achieving precise palm print and palm vein identification. Attached Figure Description

[0085] Figure 1 This is a flowchart of the palm biometric fusion and recognition method of this application;

[0086] Figure 2 Example image showing the segmentation of the effective palmprint region in a palmprint image;

[0087] Figure 3 for Figure 2 A diagram after rotation;

[0088] Figure 4 Example images of palm prints, palm veins, and the intersections of the main palm print lines and palm vein vessels;

[0089] Figure 5 Example image showing contrast enhancement operation performed on the effective area of ​​palm print;

[0090] Figure 6 An example of a palm print feature dot map;

[0091] Figure 7 An example of a palm print main line diagram;

[0092] Figure 8This is an example image of the refined palm print main lines;

[0093] Figure 9 for Figure 8 A schematic diagram after inversion;

[0094] Figure 10 Example diagram showing the intersection of the main palm lines and palmar veins;

[0095] Figure 11 This is a schematic diagram showing the orientation of the feature points to be identified.

[0096] Figure 12 This is a schematic diagram of the palm biometric fusion recognition device of this application. Detailed Implementation

[0097] To make the technical problems, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0098] Biometric multimodal fusion recognition uses a combination of multiple biometric types, sensors, samples, instances, and / or algorithms to obtain a specific biometric recognition result. Multimodal recognition fusion is divided into four levels: sample-level fusion, feature-level fusion, score-level fusion, and decision-level fusion.

[0099] Existing palmprint and palm vein fusion recognition technologies are essentially simple superposition and fusion methods. Taking feature-level fusion as an example, palmprint and palm vein recognition each provide a set of extracted features, and the fusion process combines these features into a single feature set or feature vector for comparison and decision-making. This fusion simply superimposes the features of palmprint and palm vein without considering the correlation and dependency between these features. While this approach may outperform single palmprint and single palm vein recognition to some extent, it does not significantly improve the accuracy and reliability of biometric recognition.

[0100] To address the aforementioned problems, embodiments of this application provide a palm biometric fusion recognition method, such as... Figure 1 As shown, the method includes:

[0101] S100: Simultaneously acquire palm print and palm vein images of the same hand of the same user.

[0102] For example, the acquisition device has the function of acquiring visible light palm print and near-infrared palm vein images. Simultaneous acquisition can ensure that the palm position, posture and contour of the two images are consistent in the same common area, and the two images can be processed in the same coordinate system.

[0103] S200: Obtain the effective area of ​​palm print and the effective area of ​​palm vein based on palm print image and palm vein image.

[0104] This step is used to preprocess the palmprint and palm vein images. The purpose of preprocessing is to register, normalize, and segment the effective region of the palmprint (or palm vein) images to reduce the impact of differences in palm size among different people and the translation and rotation that are unavoidable during the palm acquisition process on recognition. Finally, the effective regions of palmprint and palm vein are obtained for subsequent palmprint and palm vein extraction.

[0105] To ensure that the extracted palm prints and palm veins can be processed in the same coordinate system, the effective areas of the palm prints and palm veins should be in the same position, that is, the effective areas of the palm prints and palm veins should be in the same position on the user's palm.

[0106] S300: Extract the main palm print line based on the effective area of ​​the palm print.

[0107] The main palm lines are the deepest and thickest lines on the palm, distributed according to the location of the extensor and adductor muscle groups in different parts of the palm, and are also known as the palm flexor lines. They are generally composed of the distal palmar lines, thenar lines, and medial palmar lines.

[0108] Specifically, in the palmprint image of this application, straight lines or smooth curves with a width of 3 to 5 pixels and a length greater than 15 pixels are referred to as palmprint line features, which are mainly composed of three main palmprint lines. The image of the main palmprint lines can be extracted using a palmprint extraction algorithm.

[0109] S400: Palm vein vessels are extracted based on the effective area of ​​the palm vein.

[0110] Palm veins are the interwoven veins beneath the skin of the palm, forming a textured pattern. Hemoglobin in human blood has a high absorption coefficient for near-infrared light. When the palm is illuminated with near-infrared light of a specific wavelength, the light is absorbed by the hemoglobin in the veins, appearing as dark vein patterns. The area around the veins, lacking hemoglobin, appears bright. Palm vein extraction algorithms can be used to extract images of the palm veins.

[0111] S500: Under the same coordinate system, extract the intersection points of the palm print main line and the palm vein vessels to obtain the set of feature points to be identified.

[0112] Research has shown that the distribution of palm print lines and palm veins is relatively fixed. Utilizing this characteristic, in the embodiments of this application, the intersection points of palm print lines and palm veins are used as feature points in the same common area of ​​the palm to form a set of feature points to be identified. Figure 4 A schematic diagram is provided, showing the palm print image, palm vein image, and the intersection of the main palm print line and the palm vein.

[0113] This application achieves feature fusion of palm prints and palm veins by using the set of feature points formed by the intersection of the palm print main line and the palm veins. This fusion is essentially a feature-level fusion of multimodal recognition fusion, meaning that fusion is performed during the feature extraction stage, and then the fused features are compared. The feature extraction stage is a crucial step in the biometric system processing flow, and feature fusion at this stage plays a vital role in improving the system's recognition rate, even more significantly than score fusion or decision fusion.

[0114] This application utilizes the dependency relationship between palmprint and palm vein features to achieve the fusion of palmprint and palm vein features for individual biometric registration and identification. The resulting set of feature points to be identified reflects not only palmprint features but also palm vein features, as well as the correlation and dependency between palmprints and palm veins. It possesses the characteristics of prominent, stable, and interference-resistant palmprint features, the characteristics of concealed and interference-resistant palm veins, and the correlation and dependency between the relative positional distribution of palmprints and palm veins.

[0115] Identification is performed based on the set of feature points to be identified. This method combines the advantages of palmprint recognition, which has a thicker and more stable main line compared to other features, and the advantages of palm vein recognition, which uses near-infrared imaging of blood vessels and is less susceptible to interference. It also leverages the correlation and dependence between the relative positions of palmprints and palm veins to significantly improve the accuracy and reliability of biometric identification, thus achieving precise identification of palmprints and palm veins.

[0116] S600: Matches the set of feature points to be identified with the pre-registered template feature point set to achieve palm biometric fusion recognition.

[0117] A typical biometric identification system mainly consists of two parts: biometric registration and verification (1:1 comparison) / identification (1:N comparison). The registration process includes biometric sample acquisition, sample quality assessment, template creation, and template storage. The verification / identification process includes biometric sample acquisition, sample quality assessment, feature extraction, feature comparison, and decision-making.

[0118] The aforementioned S100 to S600 of this application describe the verification / identification process, which obtains a set of feature points to be identified and matches it with a pre-registered template feature point set. The pre-registered template feature point set is constructed in the same way as the feature point set to be identified, except that the template feature point set is pre-registered and stored before verification / identification. Therefore, the construction method of the template feature point set will not be described in detail.

[0119] This application leverages the relatively fixed distribution of palm print main lines and palm veins, using the intersections of these lines as feature points to construct a feature point set for registration and identification. This achieves feature-level fusion of palm print and palm vein characteristics, and compares the fused feature point set to realize palm biometric fusion identification. This application combines the advantages of palm print recognition (larger, more prominent, stable, and less susceptible to interference compared to other features) with palm vein recognition (concealed near-infrared imaging of blood vessels, less susceptible to interference) and the correlation and dependence between the relative positions of palm prints and palm veins, significantly improving the accuracy and reliability of biometric identification and achieving precise palm print and palm vein identification.

[0120] This application does not limit the specific implementation method for obtaining the effective area of ​​palm print and the effective area of ​​palm vein. The following uses a palm print image as an example for illustration.

[0121] 1) If the palm print image is a visible light color image, then perform grayscale processing.

[0122] 2) The grayscale palm print image is Gaussian smoothed and then binarized.

[0123] 3) The palm contour is extracted using an edge detection algorithm. This application may use the Sobel edge detection algorithm, the Canny edge detection algorithm, or other edge detection algorithms.

[0124] 4) Locate key points and establish a coordinate system to segment the effective palm print area:

[0125] Key point P1 is the intersection of the gaps between the index and middle fingers (the point closest to the palm side), and key point P2 is the intersection of the gaps between the ring and little fingers (the point closest to the palm side).

[0126] Establish a rectangular coordinate system with the midpoint of P1 and P2 as the origin, the line passing through P1 and P2 as the x-axis (the positive direction of the x-axis is to the right), and the line passing through the origin and perpendicular to the x-axis as the y-axis (the positive direction of the y-axis is downward).

[0127] Draw a line segment ab parallel to the x-axis with the midpoint (0, 10). The length of ab is the length of the extension of line segment P1P2 to both ends of line segment P1P2 by 20%. Using ab as one side of a square, draw a square region towards the palm (positive y-axis direction), which is the region of interest (ROI) of the image. This also determines the coordinates and side lengths of the four vertex corners of the square. Figure 2 As shown.

[0128] by Figure 2 Rotate the image around its origin O, making the x-axis horizontal and the y-axis vertical, as shown below. Figure 3 As shown, the area inside the rotated square is the effective area of ​​the palm print.

[0129] This completes the registration and normalization of the palmprint image. Since the palmprint image and the palm vein image are synchronized, registration and normalization can be performed according to the registration and normalization parameters of the palmprint image to obtain the effective region of palm veins (ROI).

[0130] 5) If necessary, the normalized image can also be enhanced.

[0131] This application does not limit the specific method for extracting the palmprint main line. The following is an example illustrating this method, which uses a three-step process of enhancement, segmentation, and denoising to extract the palmprint image. Specifically, the steps include:

[0132] S310: Perform contrast enhancement operation on the effective area of ​​the palm print to obtain an enhanced image.

[0133] Since the grayscale distinction between palm prints and the palm is not obvious, a contrast enhancement operation is used to enhance the image, including the following steps:

[0134] 1. Convolve the effective area of ​​the palm print using the set filter template to obtain the response image.

[0135] Wherein, the filter template is

[0136] 2. Then, the initial enhanced image is calculated using the following formula.

[0137] I L (i,j)=I0(i,j)-c L (i,j)

[0138] Among them, I L (i,j), c L (i,j) and I0(i,j) represent the initial enhanced image I, respectively. L Response image c LThe grayscale value of the pixel at coordinates (i,j) of the effective palmprint region I0, i = 1, 2, ..., M, j = 1, 2, ..., N, where M and N represent the width and height of the effective palmprint region, respectively.

[0139] 3. Finally, median filtering is applied to the initial enhanced image to obtain the enhanced image.

[0140] The initial enhanced image after contrast enhancement, although the main palm print lines are more obvious, also contains a lot of small noise. Therefore, median filtering and other methods can be used to remove the small noise to obtain the enhanced image.

[0141] Median filtering utilizes a window template with an odd number of points. The template is slid from one end of the image, aligning its center with a specific pixel location. The grayscale values ​​of the corresponding pixels under the template are read, arranged in ascending order, and the middle value is identified and assigned to the pixel at the template's center. The size of the window template can be set; this application uses a 3x3 template.

[0142] Figure 5 Example images of the original palm print effective area, the initial enhanced image, and the enhanced image are provided.

[0143] S320: Extract palmprint feature points from the enhanced image to obtain a palmprint feature point map.

[0144] Palmprint characteristics refer to the inherent properties and features of palmprints, reflecting information such as their shape. Palmprint characteristic points are points that reflect the shape of the palmprint; by extracting these points, the main lines of the palmprint can be further obtained. In this step, the palmprint image can be segmented using a local grayscale segmentation method to extract palmprint characteristic points, thereby obtaining a palmprint characteristic point map. Specifically, the steps are as follows:

[0145] 1. Divide the enhanced image into non-overlapping sub-image blocks according to the set size, and calculate the average gray value of each sub-image block.

[0146] Suppose the enhanced image is divided into sub-image blocks of size ms*ns (e.g., 32*32), and the mean grayscale value E of each sub-image block is... s The calculation formula is as follows:

[0147]

[0148] I(p,q) represents the gray value of the pixel at coordinate (p,q) of the sub-image patch, s = 1, 2, ..., NS, where NS represents the total number of sub-image patches.

[0149] 2. The palmprint characteristic point index is calculated using the local gray-level squared difference method and the following formula.

[0150] I′(i,j)=I(i,j) 2 -(ηE s ) 2

[0151] Where I(i,j) represents the gray value of the pixel at coordinate (i,j) of the enhanced image I, E s I represents the grayscale mean of the sub-image block to which I(i,j) belongs, η represents the set weight factor, and I′(i,j) represents the palmprint feature point index of the pixel at coordinate (i,j).

[0152] This application uses the local gray-level squared difference method to distinguish palm prints from the background. Specifically, it calculates the squared difference between each pixel and the mean to distinguish the main line pixels. In the calculation process, a weight factor η is given. In one example, η = 0.93.

[0153] 3. Select the pixels with palmprint feature point indices less than or equal to zero as palmprint feature points, set the grayscale value of the palmprint feature points to 1, and set the grayscale value of other pixels to 0 to obtain the palmprint feature point map.

[0154] The palmprint feature point map is denoted as c1(i,j), and the calculation formula is as follows:

[0155]

[0156] An example of the obtained palmprint feature point map is as follows: Figure 6 As shown.

[0157] S330: Generate palm print main lines based on the palm print characteristic point map to obtain the palm print main line map.

[0158] This step uses the Sobel operator and a range template (circular binary template) to estimate the principal line range, thereby removing noise outside the principal line. Finally, morphological operations and morphological thinning are used to obtain palmprint line features with a single pixel width. This includes the following steps:

[0159] 1. The range of the palmprint feature point map c1(i,j) is estimated by the Sobel operator, and the response image H(i,j) is segmented by setting a threshold Tsl. When H(i,j)>Tsl, it is the main line point of the palmprint.

[0160] 2. Using the following formula, with the set circular binary template g... n The image H(i,j) processed by the Sobel operator is traversed and searched, and the circular binary template g is changed. n The radius n is iterated to remove noise and obtain a denoised image.

[0161]

[0162] Where H1(i,j) represents the gray value of the pixel at coordinate (i,j) of the denoised image H1, g n (i c ,j c ) represents the center pixel (i) of the circular binary template. c ,j c The grayscale value of ), where H represents the image after processing by the Sobel operator, sum(g) n (H)) represents the sum of the gray levels of all pixels in the area covered by the circular binary template H, and m0 is the number of isolated noises.

[0163] 3. Then, the denoised image H1 is still treated using a circular binary template g with radius m. m Range estimation is performed to obtain the palm print main line map.

[0164] Taking m=5 as an example, the calculation formula is as follows:

[0165]

[0166] An example of the obtained palm print master line diagram is as follows: Figure 7 As shown.

[0167] S340: The palm print main line image is refined through morphological operations to obtain a palm print main line with a single pixel width.

[0168] During the refinement process, the palm print main lines are first appropriately expanded, and local palm print breakpoints are connected. Then, an opening operation is performed to remove redundant connections after expansion. Finally, morphological refinement is used to extract the skeleton and delete small branches. The refined palm print main lines are shown below. Figure 8 As shown, invert it, as... Figure 9 As shown.

[0169] At this point, the main palm lines have been extracted. The main lines can then be marked. The main palm lines closest to the four fingers (i.e., the distal palm lines) are marked as 1, the main palm lines in the middle (i.e., the central palm lines) are marked as 2, and the main palm lines closest to the wrist (i.e., the thenar eminence lines) are marked as 3.

[0170] This application does not limit the specific method for extracting palm veins. The following is an example, which includes:

[0171] S410: Perform median filtering on the effective area of ​​the palm veins to remove noise. A 3x3 window template can be used for median filtering here.

[0172] S420: Enhances the contrast of the image after median filtering.

[0173] Assuming the original image f(x, y) has a grayscale range of [a, b], and if the grayscale range of image g(x, y) is expanded to [c, d] after a linear transformation, then the following transformation can be used:

[0174]

[0175] After the above grayscale linear transformation, the dynamic range of the pixels increases, the contrast of the image is expanded, and the image becomes clearer, more delicate and easier to recognize.

[0176] S430: Binarize the contrast-enhanced image according to the set binarization threshold.

[0177] The palm vein images processed above are all 8-bit grayscale images, requiring binarization. Binarization transforms the palm vein image from a grayscale image into a binary image with only black and white. The principle is to find a suitable threshold and compare the grayscale value of each point in the image with that threshold. Pixels with a grayscale value less than the threshold are set to 0, and pixels with a grayscale value greater than the threshold are set to 1. Let the points in the contrast-enhanced image be represented by I(x, y), and the points in the binarized image be represented by I'(x, y). The predefined threshold is Th. The entire binarization transformation process is shown in the formula:

[0178]

[0179] S430: Thin the binarized image to obtain palm veins with a single pixel width.

[0180] Image thinning refers to the process of extracting a skeleton of only one pixel width as quickly as possible while preserving the original image's topological structure. The specific thinning process is as follows:

[0181] For any pixel, create a 3×3 template (P1 to P9 arranged in order), which contains 8 adjacent pixels, as shown below.

[0182]

[0183] In the binarized image, blood vessel points are 1, background points are 0, and the template center point P5 is 1, while at least one of the eight points in its neighborhood is 0. The following condition is then applied:

[0184] a) 2 ≤ N ≤ 6;

[0185] b) S = 1;

[0186] c) P2×P4×P6=0 and P4×P6×P8=0

[0187] Alternatively, P2×P4×P8=0 and P2×P6×P8=0.

[0188] Where N is the number of non-zero points in the neighborhood, and S is the number of times points P1...P4 and P6...P9 change from 0 to 1. If all the above conditions are met, then P5 is deleted. This process is repeated until there are no more points to delete. At this point, the algorithm ends, generating a region skeleton. The image is refined to the minimum connected lines without discontinuities, i.e., palm veins with a width of one pixel.

[0189] As an improvement to the embodiments of this application, the aforementioned S500 includes:

[0190] S510: In the same coordinate system, find the point on the palm print main line and the palm vein with the same coordinates, and use the intersection point as the feature point to be identified.

[0191] This process is equivalent to feature fusion, and the intersection points are the fused feature points. An example of an intersection point is shown below. Figure 10 As shown.

[0192] S520: On the same palm print line, connect each feature point to be identified in sequence from one end of the palm print line to the other (e.g., from left to right).

[0193] Example of connecting the various feature points to be identified: Figure 11 As shown.

[0194] S530: For any feature point to be identified, calculate the angle between the vector from the feature point to the next feature point and the positive x-axis, and use it as the direction of the feature point to be identified.

[0195] Assuming the connections are made from left to right, the angle θ between the vector of each connecting line and the positive x-axis represents the direction of the feature point on the left. For example... Figure 11 As shown. The fusion feature point direction representation in this application adopts a point-to-point connection method, which not only provides the direction, but also indicates the arrangement order and constraints of the feature points.

[0196] S540: The location coordinates, direction, and palm print main line identifier of each feature point to be identified are used to form a set of feature points to be identified; where the palm print main line identifier indicates the palm print main line to which the feature point to be identified belongs.

[0197] The set of feature points to be identified in this application constitutes palm print and palm vein fusion features, including the palm print main line identifier (main line identifier number 1 to 3), position coordinates (x, y), and direction (θ).

[0198] Based on the above set of feature points to be identified, hand biometric fusion recognition is performed using the following method:

[0199] S610: Determine the feature points to be identified and the template feature points belonging to the same palm print main line based on the palm print main line identifiers of the feature points to be identified contained in the feature point set to be identified and the template feature points contained in the template feature point set.

[0200] S620: Match the position coordinates and directions of the feature points to be identified and the template feature points that belong to the same palm print main line, and take the feature points to be identified and the template feature points that match in both position coordinates and direction as the matching feature point pairs.

[0201] The specific matching process is as follows:

[0202] 1. Calculate the distance information between the feature points to be identified and the template feature points based on their position coordinates, which belong to the same palm print main line.

[0203] Assume the i-th feature point to be identified is (p ix ,p iy ,p iθ The j-th template feature point is (q) jx ,q jy ,q jθ ), p ix ,p iy ,p iθ Let p represent the i-th feature point to be identified. i x-coordinate, y-coordinate, and direction, q jx ,q jy ,q jθ Let q represent the j-th template feature point respectively. j The x-coordinate, y-coordinate, and direction.

[0204] The formula for calculating the distance information dis between the i-th feature point to be identified and the j-th template feature point can be:

[0205] dis = |p ix -q jx |+|p iy -q jy |

[0206] This application does not limit the specific calculation method of the distance information; for example, it can also be Euclidean distance, etc.

[0207] 2. Compare the distance information with the preset distance range. If the distance information is within the range, the location coordinates are considered to be matched; otherwise, the location coordinates are considered to be not matched.

[0208] For example, assuming the pre-set distance range is <dis_th, if dis < dis_th, then the position coordinates match successfully.

[0209] 3. Calculate the angle difference information between the feature points to be identified and the template feature points that belong to the same palm print main line.

[0210] The formula for calculating the angle difference ang between the i-th feature point to be identified and the j-th template feature point can be:

[0211] ang = |p iθ -q jθ |

[0212] 4. Compare the angle difference information with the preset angle difference range. If the angle difference information is within the angle difference range, the direction matching is considered successful; otherwise, the direction matching is considered unsuccessful.

[0213] For example, assuming the pre-set angle difference range is <ang_th or 2π->ang_th, if ang < ang_th or ang > 2π->ang_th, then the direction matching is successful.

[0214] The aforementioned dis_th and ang_th are the set distance threshold and angle difference threshold, respectively.

[0215] 5. The feature points to be identified and the template feature points that pass both position coordinate matching and direction matching are used as matching feature point pairs.

[0216] If p i With q j If a match is found, then p i With q j As a pair of matching feature points, this process is repeated to find all feature point pairs that satisfy the above conditions as the basis for matching.

[0217] 6. Calculate the percentage of the total number of unidentified feature points and template feature points contained in the matching feature point pair to the total number of unidentified feature points contained in the unidentified feature point set and template feature points contained in the template feature point set.

[0218] 7. Determine if the proportion is greater than the set proportion threshold. If so, the matching is successful and the palm biometric fusion recognition is passed. Otherwise, the palm biometric fusion recognition fails.

[0219] For example, the percentage threshold can be set to 60%.

[0220] In addition to the palm print and palm vein fusion feature comparison mentioned above, if the system also supports single palm print and single palm vein feature comparison methods, then multiple comparison methods can be used together to perform multimodal biometric recognition.

[0221] Based on this, the method of this application also includes:

[0222] S700: Palm print recognition and palm vein recognition are performed based on the effective area of ​​palm print and the effective area of ​​palm vein, respectively.

[0223] S800: Performs multimodal biometric recognition based on palm biometric fusion recognition, palm print recognition, and palm vein recognition.

[0224] This multimodal biometric recognition can be either score-level fusion or decision-level fusion. For example, in score-level fusion, the fusion score = palm vein comparison score * palm print and palm vein fusion feature comparison score, which serves as the basis for recognition. During calculation, each comparison score is normalized to [0, 1].

[0225] This application also provides a palm biometric fusion recognition device, such as... Figure 12 As shown, the device includes:

[0226] The image acquisition module 100 is used to simultaneously acquire palm print images and palm vein images of the same hand of the same user.

[0227] The preprocessing module 200 is used to obtain the effective area of ​​the palm print and the effective area of ​​the palm vein based on the palm print image and the palm vein image.

[0228] Among them, the effective area of ​​palm print and the effective area of ​​palm vein correspond to the same location on the user's palm.

[0229] The palmprint main line extraction module 300 is used to extract the palmprint main line based on the effective area of ​​the palmprint.

[0230] The palm vein extraction module 400 is used to extract palm veins based on the effective area of ​​the palm vein.

[0231] The feature fusion module 500 is used to extract the intersection points of the palm print main line and the palm vein vessels in the same coordinate system to obtain the set of feature points to be identified.

[0232] The recognition module 600 is used to match the set of feature points to be recognized with the pre-registered template feature point set to achieve palm biometric fusion recognition.

[0233] This application leverages the relatively fixed distribution of palm print main lines and palm veins, using the intersections of these lines as feature points to construct a feature point set for registration and identification. This achieves feature-level fusion of palm print and palm vein characteristics, and compares the fused feature point set to realize palm biometric fusion identification. This application combines the advantages of palm print recognition (larger, more prominent, stable, and less susceptible to interference compared to other features) with palm vein recognition (concealed near-infrared imaging of blood vessels, less susceptible to interference) and the correlation and dependence between the relative positions of palm prints and palm veins, significantly improving the accuracy and reliability of biometric identification and achieving precise palm print and palm vein identification.

[0234] As an improvement to the embodiments of this application, the aforementioned feature fusion module includes:

[0235] The feature point determination unit is used to find points with the same coordinates on the palm print main line and the palm vein in the same coordinate system, and obtain the intersection points as the feature points to be identified.

[0236] The feature point connection unit is used to connect each feature point to be identified sequentially along the same palm print line, from one end of the palm print line to the other.

[0237] The direction determination unit is used to calculate the angle between the vector from the feature point to the next feature point and the positive x-axis for any feature point to be identified, and use this angle as the direction of the feature point to be identified.

[0238] The feature point set determination unit is used to form a set of feature points to be identified by taking the position coordinates, direction and palm print main line identifier of each feature point to be identified; wherein, the palm print main line identifier indicates the palm print main line to which the feature point to be identified belongs.

[0239] Accordingly, the recognition module includes:

[0240] The same main line feature point determination unit is used to determine the feature points to be identified and the template feature points belonging to the same palm print main line based on the palm print main line identifier of the feature points to be identified contained in the feature point set to be identified and the palm print main line identifier of the template feature points contained in the template feature point set.

[0241] The position and orientation matching unit is used to match the position coordinates and orientation of the feature points to be identified and the template feature points that belong to the same palm print main line. The feature points to be identified and the template feature points that are matched in both position coordinates and orientation are used as the matching feature point pairs.

[0242] The proportion calculation unit is used to calculate the proportion of the total number of unidentified feature points and template feature points contained in the matching feature point pair to the total number of unidentified feature points contained in the unidentified feature point set and template feature points contained in the template feature point set.

[0243] The judgment unit is used to determine whether the proportion is greater than the set proportion threshold. If it is, the palm biometric fusion recognition passes; otherwise, the palm biometric fusion recognition fails.

[0244] Furthermore, the aforementioned position and orientation matching unit includes:

[0245] The distance calculation subunit is used to calculate the distance information between the feature points to be identified and the template feature points based on their position coordinates, which belong to the same palm print main line.

[0246] The location matching subunit is used to compare the distance information with a pre-set distance range. If the distance information is within the distance range, the location coordinates are considered to be matched successfully; otherwise, the location coordinates are considered to be matched unsuccessfully.

[0247] The angle difference calculation subunit is used to calculate the angle difference information between the feature points to be identified and the template feature points based on the directions of the feature points to be identified and the template feature points belonging to the same palm print main line.

[0248] The direction matching subunit is used to compare the angle difference information with a preset angle difference range. If the angle difference information is within the angle difference range, the direction matching is judged to be successful; otherwise, the direction matching is judged to be unsuccessful.

[0249] The matching feature point pair determination sub-unit is used to take the feature points to be identified and the template feature points that have passed both position coordinate matching and orientation matching as matching feature point pairs.

[0250] This application does not limit the specific implementation of the palm print main line extraction module and the palm vein extraction module. In one example, the palm print main line extraction module includes:

[0251] The contrast enhancement unit is used to perform contrast enhancement operations on the effective area of ​​the palm print to obtain an enhanced image.

[0252] The palmprint feature point extraction unit is used to extract palmprint feature points from the enhanced image to obtain a palmprint feature point map.

[0253] The palmprint main line extraction unit is used to generate palmprint main lines based on the palmprint characteristic point map, thus obtaining the palmprint main line map.

[0254] The palmprint main line refinement unit is used to refine the palmprint main line image to obtain a palmprint main line with a single pixel width.

[0255] In another example, the palm vein extraction module includes:

[0256] The filtering unit is used to perform median filtering on the effective area of ​​the palm vein.

[0257] The contrast enhancement unit is used to enhance the contrast of the image after median filtering.

[0258] The binarization unit is used to binarize the contrast-enhanced image according to the set binarization threshold.

[0259] The palm vein refinement unit is used to refine the binarized image to obtain the palm vein.

[0260] As another improvement to the embodiments of this application, the device further includes:

[0261] The palmprint and palm vein recognition modules are used to perform palmprint recognition and palm vein recognition based on the effective areas of palmprint and palm vein, respectively.

[0262] The multimodal biometric recognition module is used for multimodal biometric recognition based on palm biometric fusion recognition, palm print recognition, and palm vein recognition.

[0263] The apparatus provided in the above embodiments corresponds one-to-one with the embodiments of the aforementioned methods in terms of its implementation principle and the resulting technical effects. For the sake of brevity, any parts of the apparatus not mentioned in the embodiments can be referred to the corresponding content in the embodiments of the aforementioned methods. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the modules and units described in this apparatus can all be referred to the corresponding processes in the embodiments of the aforementioned methods, and will not be repeated here.

[0264] The palm biometric fusion recognition method described in the above embodiments of this application can implement business logic through a computer program and record it on a storage medium. This storage medium can be read and executed by a computer, achieving the effects of the scheme described in the method embodiments of this specification. Therefore, this application also provides a computer-readable storage medium for palm biometric fusion recognition, including a memory for storing processor-executable instructions. When executed by a processor, these instructions implement the steps of the palm biometric fusion recognition method of the aforementioned embodiments.

[0265] The storage medium may include a physical device for storing information, typically digitizing the information and then storing it using electrical, magnetic, or optical methods. The storage medium may include: devices that store information using electrical energy, such as various types of memory, like RAM and ROM; devices that store information using magnetic energy, such as hard disks, floppy disks, magnetic tapes, magnetic core memory, bubble memory, and USB flash drives; and devices that store information using optical methods, such as CDs or DVDs. Of course, there are other readable storage media, such as quantum memories and graphene memories.

[0266] The storage medium described above may also include other implementation methods according to the description of the method embodiments. The implementation principle and technical effects of this embodiment are the same as those of the foregoing method embodiments. For details, please refer to the description of the relevant method embodiments, which will not be repeated here.

[0267] This application also provides a device for palm biometric fusion recognition. The device can be a standalone computer, or it can include an actual operating device that uses one or more of the methods or embodiments described in this specification. The palm biometric fusion recognition device may include at least one processor and a memory storing computer-executable instructions. When the processor executes the instructions, it implements the steps of any one or more of the palm biometric fusion recognition methods described above.

[0268] The device described above may also include other implementation methods according to the method embodiments. The implementation principle and technical effects of this embodiment are the same as those of the foregoing method embodiments. For details, please refer to the description of the relevant method embodiments, which will not be repeated here.

[0269] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for palm biometric fusion and recognition, characterized in that, The method includes: Simultaneously acquire palm print and palm vein images of the same hand of the same user; Based on the palm print image and the palm vein image, obtain the effective area of ​​the palm print and the effective area of ​​the palm vein. The effective area of ​​the palm print and the effective area of ​​the palm vein are located at the same position on the user's palm. The main palm print line is extracted based on the effective area of ​​the palm print. The palm vein vessels were extracted based on the effective area of ​​the palm vein. In the same coordinate system, the intersection points of the palm print main line and the palm vein are extracted to obtain the set of feature points to be identified; The set of feature points to be identified is matched with a pre-registered set of template feature points to achieve palm biometric fusion recognition.

2. The palm biometric fusion and recognition method according to claim 1, characterized in that, In the same coordinate system, the intersection points of the palm print main line and the palm vein are extracted to obtain the set of feature points to be identified, including: In the same coordinate system, find the point on the palm print main line with the same coordinate as the palm vein, and obtain the intersection point as the feature point to be identified; On the same palm print line, connect each feature point to be identified in sequence from one end of the palm print line to the other. For any feature point to be identified, calculate the angle between the vector from the feature point to the next feature point and the positive x-axis, and use it as the direction of the feature point to be identified; The set of feature points to be identified is formed by the position coordinates, orientation, and palm print main line identifier of each feature point to be identified; wherein, the palm print main line identifier represents the palm print main line to which the feature point to be identified belongs.

3. The palm biometric fusion and recognition method according to claim 2, characterized in that, The step of matching the set of feature points to be identified with a pre-registered template feature point set to achieve palm biometric fusion recognition includes: Based on the palmprint main line identifiers of the feature points to be identified contained in the set of feature points to be identified and the palmprint main line identifiers of the template feature points contained in the set of template feature points, the feature points to be identified and the template feature points belonging to the same palmprint main line are determined. The position coordinates and directions of the feature points to be identified and the template feature points belonging to the same palm print main line are matched respectively. The feature points to be identified and the template feature points that are matched in both position coordinates and direction are taken as the matching feature point pairs. Calculate the percentage of the total number of to-be-identified feature points and template feature points included in the matching feature point pair to the total number of to-be-identified feature points included in the to-be-identified feature point set and the total number of template feature points included in the template feature point set; Determine whether the percentage is greater than the set percentage threshold. If it is, the palm biometric fusion recognition passes; otherwise, the palm biometric fusion recognition fails.

4. The palm biometric fusion and recognition method according to claim 3, characterized in that, The process of matching the position coordinates and directions of the feature points to be identified and the template feature points belonging to the same palm print main line, and taking the feature points to be identified and the template feature points that match in both position coordinates and direction as a pair of matched feature points, includes: The distance information between the feature point to be identified and the template feature point is calculated based on the position coordinates of the feature point to be identified and the template feature point belonging to the same palm print main line. The distance information is compared with a preset distance range. If the distance information is within the distance range, the location coordinates are determined to be a successful match; otherwise, the location coordinates are determined to be a failed match. Calculate the angle difference information between the feature points to be identified and the template feature points that belong to the same palm print main line; The angle difference information is compared with a preset angle difference range. If the angle difference information is within the angle difference range, the direction matching is determined to be successful; otherwise, the direction matching is determined to be unsuccessful. The feature points to be identified and the template feature points that pass both position coordinate matching and direction matching are used as matching feature point pairs.

5. The palm biometric fusion and recognition method according to any one of claims 1-4, characterized in that, The step of extracting the palm print main line based on the effective area of ​​the palm print includes: A contrast enhancement operation is performed on the effective area of ​​the palm print to obtain an enhanced image; Palmprint feature points are extracted from the enhanced image to obtain a palmprint feature point map; Based on the palmprint feature point map, generate the palmprint main line to obtain the palmprint main line map; The palmprint main line image is refined to obtain a palmprint main line with a single pixel width.

6. The palm biometric fusion and recognition method according to any one of claims 1-4, characterized in that, The step of extracting palm vein vessels based on the effective area of ​​the palm vein includes: Median filtering is applied to the effective region of the palm vein; Enhance the contrast of the image after median filtering; The contrast-enhanced image is binarized according to the set binarization threshold. The binarized image is thinned to obtain the palmar vein.

7. The palm biometric fusion and recognition method according to any one of claims 1-4, characterized in that, The method further includes: Palm print recognition and palm vein recognition are performed based on the effective area of ​​the palm print and the effective area of ​​the palm vein, respectively. Multimodal biometric recognition is performed based on the fusion recognition of palm biometric features, the palm print recognition, and the palm vein recognition.

8. A palm biometric fusion recognition device, characterized in that, The device includes: The image acquisition module is used to simultaneously acquire palm print images and palm vein images of the same hand of the same user; The preprocessing module is used to obtain the effective area of ​​the palm print and the effective area of ​​the palm vein based on the palm print image and the palm vein image. The effective area of ​​the palm print and the effective area of ​​the palm vein are located at the same position on the user's palm. The palmprint main line extraction module is used to extract the palmprint main line based on the effective area of ​​the palmprint. A palm vein extraction module is used to extract palm veins based on the effective area of ​​the palm vein. The feature fusion module is used to extract the intersection points of the palm print main line and the palm vein vessels in the same coordinate system to obtain the set of feature points to be identified. The identification module is used to match the set of feature points to be identified with a pre-registered set of template feature points to achieve palm biometric fusion identification.

9. A computer-readable storage medium for palm biometric fusion recognition, characterized in that, It includes a memory for storing processor-executable instructions, which, when executed by the processor, implement the steps of the palm biometric fusion recognition method according to any one of claims 1-7.

10. A device for palm biometric fusion recognition, characterized in that, It includes at least one processor and a memory storing computer-executable instructions, wherein the processor executes the instructions to implement the steps of the palm biometric fusion recognition method according to any one of claims 1-7.

Images