WO2019075601A1

WO2019075601A1 - Palm vein recognition method and device

Info

Publication number: WO2019075601A1
Application number: PCT/CN2017/106307
Authority: WO
Inventors: 陈书楷; 程雪
Original assignee: 厦门中控智慧信息技术有限公司
Priority date: 2017-10-16
Filing date: 2017-10-16
Publication date: 2019-04-25
Also published as: CN107980140A; CN107980140B

Abstract

A palm vein recognition method and device, being used for extracting a certain number of features from each part of a region of interest of a palm, improving the recognition rate of the palm vein. The method of the embodiments of the present application comprises: acquiring a target palm vein image of a user (101); extracting, from the target palm vein image, an image of a region of interest (ROI) (102); dividing the image of the ROI into at least two sub-regions (103); using a preset algorithm to extract target feature points from each of the sub-regions (104); performing feature comparison on the extracted target feature points and preset feature points, so as to obtain matching point pairs (105); determining whether each pair of the matching point pairs is a true match (106); if it is a true match, retaining the matching point pair (107); if it is a false match, removing the matching point pair (108).

Description

Palm vein identification method and device

Technical field

The present application relates to the field of biometrics, and in particular, to a method and a device for identifying a palm vein.

Background technique

Palm vein recognition is the use of human hemoglobin to absorb near-infrared light through the veins, collecting vein images under the palm skin and extracting them as biometric features. The palm vein is non-contact, it is more hygienic and suitable for public use. At the same time, the application of the palm is also more natural, making it easier for users to accept. Palm vein recognition compared with other biometric technologies such as fingerprints and irises, because of the three characteristics of living body recognition, internal features and non-contact, the palm vein is extremely difficult to copy forgery, ensuring the palm vein of the user. Features are difficult to forge, so the palm vein recognition system has a high level of safety and is especially suitable for use in places with high safety requirements.

At present, the recognition method of the palm vein is based on the overall subspace learning method, that is, the whole palm vein is taken as a global description, and the palm vein image is projected to the subspace to extract the feature vector, for example, the palm vein matching is performed by the feature recognition method. The Laplacian palm image with the palmprint and palm vein images is globally matched, and finally the local structural features of the palm vein image are extracted.

For the holistic subspace learning method, due to the distribution characteristics of the palm vein, the feature points are concentrated in the palm area near the big fish, and there are fewer feature points in the palm area near the small fish, which may result in the algorithm due to incomplete feature information. The recognition rate has dropped.

Summary of the invention

The embodiment of the present application provides a palm vein recognition method and device for extracting a certain number of features in each part of the palm region of interest, thereby improving the recognition rate of the palm vein.

A first aspect of the present application provides a method for identifying a palm vein, comprising: acquiring a target palm vein image of a user; extracting an image of the ROI of the region of interest from the target palm vein image; and dividing the image of the ROI into at least Two sub-regions; extracting target feature points from each of the sub-regions by using a preset algorithm; comparing the extracted target feature points with preset feature points to obtain matching point pairs; determining each pair of said Whether the matching point pair is a true match; if it is a true match, the matching point pair is retained; If it is a false match, the matching point pair is eliminated.

In a possible design, in a first implementation manner of the first aspect of the embodiments of the present application, the determining whether each pair of the matching point pairs is a true match comprises: dividing the image of the ROI into the same specifications G*G grids, the G is a positive integer greater than 1, and each pair of the matching point pairs is mapped to corresponding positions of the G*G grids to obtain a preset feature point lattice image and a target feature point a grid image; determining a first grid and a second grid having the largest number of identical matching point pairs as matching lattice pairs, the first lattice is located in the preset feature point lattice image, and the second lattice is located in the target feature Point grid image; determining whether each pair of matching points belongs to a corresponding matching lattice pair; if the matching point pair belongs to a corresponding matching lattice pair, determining that the matching point pair is a true match; if the matching point pair is If it does not belong to the corresponding matching pair, it is determined that the matching point pair is a false match.

In a possible implementation, in a second implementation manner of the first aspect of the embodiments of the present application, determining whether each pair of the matching point pairs belongs to a corresponding matching cell pair comprises: calculating the first lattice a threshold value and a score value of the second grid; determining whether the score value is greater than the threshold; if yes, determining that the matching point pair belongs to a corresponding matching lattice pair; if not, determining that the matching point pair does not belong to Corresponding match pairs.

In a possible design, in a third implementation manner of the first aspect of the embodiments of the present application, the adopting a preset algorithm to extract target feature points from each of the sub-regions includes: according to each of the sub-regions The region adjusts a sampling threshold, the sampling threshold is used to determine the target feature point; the target point in each of the sub-regions is determined as a target feature point, and the parameter value of the target point is greater than the sampling threshold.

In a possible design, in a fourth implementation manner of the first aspect of the embodiments, the preset algorithm is a direction descriptor ORB algorithm, a scale invariant feature transform SIFT algorithm, or a fast robust feature SURF algorithm. Any of the algorithms.

The second aspect of the present application provides an apparatus for identifying a palm vein, comprising: an acquiring unit, configured to acquire a target palm vein image of a user; and a first extracting unit, configured to extract a region of interest ROI from the target palm vein image And a dividing unit, configured to divide the image of the ROI into at least two sub-regions; and a second extracting unit, configured to extract a target feature point from each of the sub-regions by using a preset algorithm; And comparing the extracted target feature points with the preset feature points to obtain a matching point pair; the determining unit is configured to determine whether each pair of the matching point pairs is a true match; if the matching unit is a true match , for retaining the matching point pair; the culling unit, if it is a false match, is used for culling The matching point pairs.

In a possible design, in a first implementation manner of the second aspect of the embodiment of the present application, the determining unit includes: a processing module, configured to divide the image of the ROI into G*G pieces with the same specifications a grid, wherein G is a positive integer greater than 1, and each pair of the matching point pairs is mapped to a corresponding position of the G*G grids to obtain a preset feature point grid image and a target feature point grid image; a determining module, configured to determine a first lattice and a second lattice having the largest number of identical matching point pairs as a matching lattice pair, the first lattice is located in the preset feature point lattice image, and the second lattice is located in the a target feature point grid image; a judging module, configured to determine whether each pair of the matching point pairs belongs to a corresponding matching lattice pair; and the second determining module, if the matching point pair belongs to the corresponding matching lattice pair, is used to determine the The matching point pair is a true match; the third determining module is configured to determine that the matching point pair is a false match if the matching point pair does not belong to the corresponding matching pair.

In a possible design, in a second implementation manner of the second aspect of the embodiment, the determining module is specifically configured to: calculate a threshold of the first grid and a score of the second grid; Determining whether the score value is greater than the threshold; if yes, determining that the matching point pair belongs to a corresponding matching lattice pair; if not, determining that the matching point pair does not belong to a corresponding matching lattice pair.

In a possible design, in a third implementation manner of the second aspect of the embodiments of the present application, the second extraction unit includes: an adjustment module, configured to adjust a sampling threshold according to each of the sub-areas, The sampling threshold is used to determine the target feature point; and a fourth determining module is configured to determine a target point in each of the sub-regions as a target feature point, where a parameter value of the target point is greater than the sampling threshold.

In a possible design, in a fourth implementation manner of the second aspect of the embodiments of the present application, the preset algorithm is a direction descriptor ORB algorithm, a scale invariant feature transform SIFT algorithm, or a fast robust feature SURF algorithm. Any of the algorithms.

A third aspect of the present application provides a device for identifying a palm vein, comprising: a memory and at least one processor, wherein the memory stores an instruction, and the memory and the at least one processor are interconnected by a line; the at least one The processor invokes the instructions in the memory to cause the palm vein identification device to perform the methods described in the various aspects above.

A fourth aspect of the present application provides a computer readable storage medium having instructions stored therein that, when executed on a computer, cause the computer to perform the methods described in the above aspects.

A fifth aspect of the present application provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform the methods described in the various aspects above.

As can be seen from the above technical solutions, the embodiments of the present application have the following advantages:

In the technical solution provided by the embodiment of the present application, the target palm vein image of the user is acquired; the image of the ROI of the region of interest is extracted from the target palm vein image; the image of the ROI is divided into at least two sub-regions; The algorithm extracts a target feature point from each of the sub-regions; compares the extracted target feature points with preset feature points to obtain a matching point pair; and determines whether each pair of the matching point pairs is a true match If it is a true match, the matching point pair is retained; if it is a false match, the matching point pair is eliminated. In the embodiment of the present application, the image of the ROI is divided into multiple sub-regions, and feature points are extracted from each sub-region, thereby ensuring that a certain number of feature points are extracted in each part of the palm region of interest, and the palm vein is improved. Recognition rate.

DRAWINGS

1 is a schematic diagram of an embodiment of a method for identifying a palm vein in an embodiment of the present application;

2 is a schematic diagram of another embodiment of a method for identifying a palm vein according to an embodiment of the present application;

3 is a schematic diagram of a position of a feature point in an ROI region according to an embodiment of the present application;

4 is a schematic diagram of distribution of feature points in a grid in the embodiment of the present application;

FIG. 5 is a schematic diagram of a comparison of matching point pairs and matching lattice pairs in the embodiment of the present application; FIG.

6 is a schematic diagram of an embodiment of a palm vein identification device according to an embodiment of the present application;

FIG. 7 is a schematic diagram of another embodiment of a palm vein identification device according to an embodiment of the present application; FIG.

FIG. 8 is a schematic diagram of another embodiment of an apparatus for identifying a palm vein according to an embodiment of the present application.

Detailed ways

The embodiment of the present application provides a palm vein recognition method and device for extracting a certain number of feature points in each part of the palm region of interest, thereby improving the recognition rate of the palm vein.

The embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

The terms "first", "second", "third", "fourth", etc. (if present) in the specification and claims of the present application and the above figures are used to distinguish similar objects without having to use To describe a specific order or order. It should be understood that the data used in this way can be interchanged where appropriate, so here The described embodiments can be implemented in a sequence other than what is illustrated or described herein. In addition, the term "comprises" or "comprises" or any variations thereof, is intended to cover a non-exclusive inclusion, for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those that are clearly listed Steps or units, but may include other steps or units not explicitly listed or inherent to such processes, methods, products or devices.

For ease of understanding, the specific process of the embodiment of the present application is described below. Referring to FIG. 1 , an embodiment of the method for identifying a palm vein in the embodiment of the present application includes:

101. Acquire a target palm vein image of the user.

The user places the palm of the hand in the scanning range of the identification device, and the identification device turns on the near-infrared light to obtain the target palm vein image of the user. Among them, since the venous blood of the palm can absorb the near-infrared light, the venous blood vessel reflects less and is darker than the surrounding, thereby forming a palm vein pattern.

It should be noted that the near-infrared light is an electromagnetic wave between visible light and mid-infrared light, and refers to an electromagnetic wave having a wavelength in the range of 780 to 2526 nm. It can be understood that in the process of recognizing the palm of the user, the palm side of the palm of the user receives near-infrared light illumination, and the palm of the user is all located within the illumination range of the identification device to ensure the integrity of the acquired target palm vein image. Sex.

102. Extract an image of the ROI of the region of interest from the target palm vein image.

An image of a region of interest (ROI) is extracted from the target palm vein image. The pixel size of the ROI area may be set to 184×184, or may be 129×129, and may be other values, which may be set according to actual needs, which is not limited herein.

103. Divide the image of the ROI into at least two sub-regions.

The image of the ROI is divided into at least two sub-regions. Specifically, the image of the ROI is equally divided into N*N sub-regions, and N is a positive integer greater than 1. Generally, when the value of N is N=2 or N=3, the distribution of the feature points is reasonable, and the value of N can also be It is other values, which are not limited here.

104. Extract a target feature point from each sub-area by using a preset algorithm.

After the image of the ROI is divided into N*N sub-regions that are equally divided, in each sub-region, the target feature points are extracted using a preset algorithm. Specifically, the direction descriptor (ORB) algorithm is used for feature extraction and representation. For different sub-regions, the sampling threshold is also different. Each sub-region can adjust the sampling threshold. The sampling threshold is used to determine whether the target point is the target feature. Point to enable the sub-region to extract feature points that satisfy the number of conditional numbers. Specifically, when the parameter value of the target point is greater than When the sampling threshold of the sub-region is determined, the target point is determined as the target feature point. When the parameter value of the target point is less than or equal to the sampling threshold of the sub-region, determining that the target point is not the target feature point, and the target point is excluded. Iterate through all the target points in the image of the ROI and extract the target feature points that meet the quantity requirements.

It can be understood that the preset algorithm may also be other feature extraction algorithms. For example, the preset algorithm may also be a scale invariant feature transform (SIFT) algorithm or a speeded-up robust feature (SURF). The algorithm may also be another feature extraction algorithm, which is not limited herein.

105. Compare the extracted target feature points with the preset feature points to obtain matching point pairs.

The extracted target feature points are compared with the preset feature points to obtain matching point pairs. Specifically, the extracted number of target feature points are compared with a database, and the database includes preset feature points and feature parameters of the preset feature points. After the target feature points are acquired, the feature parameters of each target feature point are compared with the feature parameters in the database, and one preset feature point and one target feature point with the same feature parameter are determined as matching point pairs.

It should be noted that there are a plurality of matching methods. Specifically, it may be a brute force (BF) algorithm, or may be another matching algorithm, which is not limited herein.

106. Determine whether each pair of matching points is a true match.

Determine if each pair of matching points is a true match. Specifically, each pair of matching point pairs obtained is filtered, and the matching point pairs that meet the requirements are determined to be true matches according to the screening criteria, and the matching point pairs that do not meet the requirements are false matches. If it is a true match, step 107 is performed, and if it is a false match, step 108 is performed.

107. Keep matching point pairs.

If the matching point pair is a true match, the matching point pair is retained.

108. Eliminate matching point pairs.

If the matching point pair is a false match, the matching point pair is rejected.

In the embodiment of the present application, the image of the ROI is divided into multiple sub-regions, and feature points are extracted from each sub-region, thereby ensuring that a certain number of feature points are extracted in each part of the palm region of interest, and the palm vein is improved. Recognition rate.

Referring to FIG. 2, another embodiment of the method for identifying a palm vein in the embodiment of the present application includes:

201. Acquire a target palm vein image of the user.

The user places the palm of the hand in the scanning range of the identification device, and the identification device turns on the near-infrared light to obtain Take the user's target palm vein image. Among them, since the venous blood of the palm can absorb the near-infrared light, the venous blood vessel reflects less and is darker than the surrounding, thereby forming a palm vein pattern.

202. Extract an image of the ROI of the region of interest from the target palm vein image.

203. Divide the image of the ROI into at least two sub-regions.

204. Extract a target feature point from each sub-area by using a preset algorithm.

After the image of the ROI is divided into equal parts of N*N sub-regions, in each sub-region, the target feature points are extracted by using an oriented descriptor (ORB) algorithm. Specifically, the ORB algorithm is used for feature extraction and representation. For different sub-regions, the sampling threshold is also different. Each sub-region can adjust the sampling threshold. The sampling threshold is used to determine whether the target point is the target feature point, so that the sub-region can Feature points that satisfy the number of conditional conditions are extracted. Specifically, when the parameter value of the target point is greater than the sampling threshold of the sub-region, the target point is determined as the target feature point, and when the parameter value of the target point is less than or equal to the sampling threshold of the sub-region, determining that the target point is not The target feature point excludes the target point, traverses all the target points in the image of the ROI, and extracts the target feature points satisfying the quantity requirement.

It can be understood that the algorithm for extracting features may also be other feature extraction algorithms, for example, a scale invariant feature transform (SIFT) algorithm or a speeded-up robust features (SURF) algorithm. It can also be another feature extraction algorithm, which is not limited herein.

205. Compare the extracted target feature points with the preset feature points to obtain matching point pairs.

For example, as shown in FIG. 3, after comparing the target feature points with the database by the BF algorithm, five pairs of matching points with the same parameters are obtained, and the figures are distinguished by different geometric shapes, and may also be other numbers of matches. For point pairs and geometric shapes, this example uses five pairs of matching point pairs as an example, and is not defined as a pair of matching points. In this embodiment, the feature points represented by the triangles in the preset feature point image and the feature points represented by the triangles in the target feature point image are a pair of matching point pairs, and the feature points represented by other geometric shapes are similar, and in this embodiment, there are a total of 5 pairs of matching points.

206. Acquire a preset feature point grid image and a target feature point grid image of the matching point pair.

The image of the ROI is divided into G*G grids of the same specification, G is a positive integer greater than 1, and each pair of matching point pairs is mapped to corresponding positions of G*G grids to obtain preset feature point lattice images and targets. Feature point grid image. It can be 4 x 4 grids, and G can be other values, for example, G=20.

It should be noted that, in order to avoid the matching point pair falling into the edge of the grid or even the grid line, when the grid is divided, the grid of the right graph is moved up and down and left and right by 0.5 grids respectively, so that the matching effect is optimal. At the same time, in order to adapt to the matching between the pictures with large difference in size, the target feature point lattice map will also be scaled by a certain multiple to obtain the best matching effect.

For example, as shown in FIG. 4, the image of the ROI is divided into 4×4 grids of the same specification, and the acquired feature points in the preset feature point image and the target feature point image are corresponding to 4×4 grids. Corresponding positions, a preset feature point lattice image and a target feature point lattice image are obtained.

207. Determine a matching lattice pair according to the preset feature point lattice image and the target feature point lattice image.

The first lattice and the second lattice having the largest number of identical matching point pairs are determined as matching lattice pairs, the first lattice is located in the preset feature point image, and the second lattice is located in the target feature point image.

For example, as shown in FIG. 5, the number of feature points in each grid of the preset feature point grid image in FIG. 5 and the number of feature points in each grid of the target feature point grid image are respectively counted. Pre The first lattice of the feature point lattice image includes three preset feature points, a feature point represented by one triangle, a feature point represented by a circle, and a feature point represented by a square; the second lattice of the target feature point lattice image includes Two feature points, a feature point represented by one triangle and a feature point represented by a circle; another lattice of the target feature point lattice image includes a feature point represented by a square. Because the second lattice has the same number of feature points as the first lattice, the second lattice is determined as a lattice matching the first lattice, that is, the first lattice and the second lattice are tentative matching lattice pairs, for ease of understanding, the graph The first grid and the second grid are marked with diagonal lines.

Specifically, the threshold of the first grid is calculated by centering on the first grid and the second grid of the tentative matching lattice pair respectively. The calculation threshold value is: where: ni is the average value of the number of feature points included in the first lattice and its eight-domain grid in the preset feature point lattice image in FIG. 5, which is the set empirical value; The score value, the formula for calculating the score is:, where k is the index of the second lattice and its eight neighborhood lattice, which is the lattice image of the matching point pair falling into the target feature point in the i-th lattice of the preset feature point lattice image in FIG. Corresponds to the number of the jth grid. If the score value is greater than the threshold, the first grid and the second grid in FIG. 5 are matched pairs. Traverse all feature points to determine all matching pairs.

208. Determine whether each pair of matching points belongs to a corresponding matching pair.

Determine whether each pair of matching points belongs to the corresponding matching pair. Specifically, calculating a threshold value of the first lattice and a score value of the second lattice; determining whether the score value is greater than a threshold; if yes, determining that the matching point pair belongs to the corresponding matching lattice pair, performing step 209; if not, determining the matching point pair Step 211 is performed without belonging to the corresponding matching pair. Traverse all matching point pairs in the ROI's image.

It should be noted that there is no specific order between the threshold of the first grid and the score of the second grid. The threshold of the first grid may be calculated first, and then the score of the second grid may be calculated; or the second may be calculated first. The score value of the grid is used to calculate the threshold of the first grid; the calculation may be performed at the same time, which is not limited herein.

209. Determine that the matching point pair is a true match, and retain the matching point pair.

If the matching point pair belongs to the corresponding matching pair, it is determined that the matching point pair is a true match, and the matching point pair is retained.

210. Determine that the matching point pair is a false match, and eliminate the matching point pair.

If the matching point pair does not belong to the corresponding matching pair, it is determined that the matching point pair is a false match, and the matching point pair is eliminated.

In the embodiment of the present application, the image of the ROI is divided into multiple sub-regions, and feature points are extracted from each sub-region, so that each local part is extracted to a certain number of feature points for matching, and the matched point pairs are judged. The false matching point pairs are eliminated, and the true matching point pairs are retained, which improves the recognition rate of the algorithm and further improves the recognition rate of the palm vein.

The method for identifying the palm vein in the embodiment of the present application is described above. The following describes the palm vein identification device in the embodiment of the present application. Referring to FIG. 6, an embodiment of the palm vein identification device in the embodiment of the present application includes:

The obtaining unit 601 is configured to acquire a target palm vein image of the user;

a first extracting unit 602, configured to extract an image of the ROI of the region of interest from the target palm vein image;

a dividing unit 603, configured to divide an image of the ROI into at least two sub-regions;

a second extracting unit 604, configured to extract a target feature point from each of the sub-regions by using a preset algorithm;

The matching unit 605 is configured to compare the extracted target feature points with the preset feature points to obtain a matching point pair;

The determining unit 606 is configured to determine whether each pair of the matching points is a true match;

The retaining unit 607, if it is a true match, is used to reserve the matching point pair;

The culling unit 608, if it is a false match, is used to cull the pair of matching points.

Referring to FIG. 7, another embodiment of the palm vein identification device in the embodiment of the present application includes:

The obtaining unit 701 is configured to acquire a target palm vein image of the user;

a first extracting unit 702, configured to extract an image of the ROI of the region of interest from the target palm vein image;

a dividing unit 703, configured to divide an image of the ROI into at least two sub-regions;

a second extracting unit 704, configured to extract a target feature point from each of the sub-regions by using a preset algorithm;

The matching unit 705 is configured to compare the extracted target feature points with the preset feature points Yes, get a matching pair of points;

The determining unit 706 is configured to determine whether each pair of the matching points is a true match;

The retaining unit 707, if it is a true match, is used to reserve the matching point pair;

The culling unit 708, if it is a false match, is used to cull the pair of matching points.

Optionally, the determining unit 706 may further include:

The processing module 7061 is configured to divide the image of the ROI into G*G grids of the same specification, the G is a positive integer greater than 1, and map each pair of the matching point pairs to the G*G Corresponding positions of the grid, obtaining a preset feature point grid image and a target feature point grid image;

a first determining module 7062, configured to determine a first lattice and a second lattice having the largest number of identical matching point pairs as a matching lattice pair, where the first lattice is located in the preset feature point lattice image, and the second lattice Located at the target feature point grid image;

The determining module 7063 is configured to determine whether each pair of the matching points belongs to a corresponding matching pair;

a second determining module 7064, configured to determine that the matching point pair is a true match if the matching point pair belongs to a corresponding matching cell pair;

The third determining module 7065 is configured to determine that the matching point pair is a false match if the matching point pair does not belong to the corresponding matching cell pair.

Optionally, the determining module 7063 is further specifically configured to:

Calculating a threshold of the first grid and a score of the second grid;

Determining whether the score value is greater than the threshold;

If yes, it is determined that the matching point pair belongs to the corresponding matching lattice pair; if not, it is determined that the matching point pair does not belong to the corresponding matching lattice pair.

Optionally, the second extraction unit 704 may further include:

The adjusting module 7041 is configured to adjust a sampling threshold according to each of the sub-regions, where the sampling threshold is used to determine the target feature point;

The fourth determining module 7042 is configured to determine a target point in each of the sub-areas as a target feature point, where a parameter value of the target point is greater than the sampling threshold.

In the embodiment of the present application, the image of the ROI is divided into multiple sub-regions, and feature points are extracted from each sub-region, so that each local part is extracted to a certain number of feature points for matching, and the matched point pairs are judged. , eliminating false matching point pairs, retaining true matching point pairs, improving the algorithm recognition rate, and further The step improves the recognition rate of the palm vein.

The above-mentioned FIG. 6 to FIG. 7 describe the palm vein identification device in the embodiment of the present application in detail from the perspective of the modular functional entity. The identification device of the palm vein in the embodiment of the present application is described in detail below from the perspective of hardware processing.

FIG. 8 is a schematic structural diagram of a palm vein identification device according to an embodiment of the present application. The palm vein identification device 800 may have a large difference due to different configurations or performances, and may include one or more processors (central Processing units (CPU) 801 (eg, one or more processors) and memory 809, one or more storage media 808 (eg, one or one of the Shanghai quantity storage devices) that store application 809 or data 809. Among them, the memory 809 and the storage medium 808 may be short-term storage or persistent storage. The program stored on storage medium 808 can include one or more modules (not shown), each of which can include a series of instruction operations in the identification device for the palm vein. Still further, the processor 801 can be configured to communicate with the storage medium 808 to perform a series of instruction operations in the storage medium 808 on the palm vein identification device 800.

The palm vein identification device 800 may also include one or more power sources 802, one or more wired or wireless network interfaces 803, one or more input and output interfaces 804, and/or one or more operating systems 805, such as Windows. Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that the identification device structure of the palm vein shown in FIG. 8 does not constitute a limitation of the identification device for the palm vein, and may include more or less components than those illustrated, or may combine some components. Or different parts arrangement.

A person skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the system, the device and the unit described above can refer to the corresponding process in the foregoing method embodiment, and details are not described herein again.

In the several embodiments provided by the present application, it should be understood that the disclosed system, apparatus, and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of the unit is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application, in essence or the contribution to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like, which can store program code. .

The above embodiments are only used to explain the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still The technical solutions described in the embodiments are modified, or the equivalents of the technical features are replaced by the equivalents. The modifications and substitutions of the embodiments do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

A method for identifying a palm vein, comprising:

Obtaining the target palm vein image of the user;

Extracting an image of the ROI of the region of interest from the target palm vein image;

Dividing the image of the ROI into at least two sub-regions;

Extracting target feature points from each of the sub-regions using a preset algorithm;

Comparing the extracted target feature points with the preset feature points to obtain a matching point pair;

Determining whether each pair of the matching points is a true match;

If it is a true match, the matching point pair is retained; if it is a false match, the matching point pair is eliminated.
The identification method according to claim 1, wherein the determining whether each pair of the matching point pairs is a true match comprises:

Dividing the image of the ROI into G*G grids of the same specification, the G being a positive integer greater than 1, and mapping each pair of the matching point pairs to corresponding positions of the G*G grids, Presetting a feature point lattice image and a target feature point lattice image;

Determining a first lattice and a second lattice having the largest number of identical matching point pairs as a matching lattice pair, the first lattice is located in the preset feature point lattice image, and the second lattice is located in the target feature point lattice image ;

Determining whether each pair of the matching points belongs to a corresponding matching pair;

If the matching point pair belongs to the corresponding matching lattice pair, determining that the matching point pair is a true match;

If the matching point pair does not belong to the corresponding matching cell pair, it is determined that the matching point pair is a false match.
The identification method according to claim 2, wherein the determining whether each pair of the matching points belongs to a corresponding matching lattice pair comprises:

Calculating a threshold of the first grid and a score of the second grid;

Determining whether the score value is greater than the threshold;

If yes, it is determined that the matching point pair belongs to the corresponding matching lattice pair; if not, it is determined that the matching point pair does not belong to the corresponding matching lattice pair.
The identification method according to any one of claims 1 to 3, wherein the extracting target feature points from each of the sub-areas by using a preset algorithm comprises:

Adjusting a sampling threshold according to each of the sub-regions, the sampling threshold is used to determine the target feature point;

A target point in each of the sub-regions is determined as a target feature point, and a parameter value of the target point is greater than the sampling threshold.
The identification method according to any one of claims 1 to 3, wherein the preset algorithm is any one of a direction descriptor ORB algorithm, a scale invariant feature conversion SIFT algorithm or a fast robust feature SURF algorithm. .
A device for identifying a palm vein, comprising:

An obtaining unit, configured to acquire a target palm vein image of the user;

a first extracting unit, configured to extract an image of the ROI of the region of interest from the target palm vein image;

a dividing unit, configured to divide an image of the ROI into at least two sub-regions;

a second extracting unit, configured to extract a target feature point from each of the sub-regions by using a preset algorithm;

a matching unit, configured to compare the extracted target feature points with the preset feature points to obtain a matching point pair;

a determining unit, configured to determine whether each pair of the matching points is a true match;

The reserved unit, if it is a true match, is used to reserve the matching point pair;

The culling unit, if it is a false match, is used to cull the pair of matching points.
The identification device according to claim 6, wherein the determining unit comprises:

a processing module, configured to divide an image of the ROI into G*G grids of the same specification, the G is a positive integer greater than 1, and map each pair of the matching point pairs to the G*G grids Corresponding positions, obtaining a preset feature point grid image and a target feature point grid image;

a first determining module, configured to determine a first lattice and a second lattice having the largest number of identical matching point pairs as a matching lattice pair, wherein the first lattice is located in the preset feature point lattice image, and the second lattice is located The target feature point lattice image;

a judging module, configured to determine whether each pair of the matching points belongs to a corresponding matching pair;

a second determining module, configured to determine that the matching point pair is a true match if the matching point pair belongs to a corresponding matching cell pair;

The third determining module is configured to determine that the matching point pair is a false match if the matching point pair does not belong to the corresponding matching cell pair.
The identification device according to claim 7, wherein the determining module is specifically configured to:

Calculating a threshold of the first grid and a score of the second grid;

Determining whether the score value is greater than the threshold;

If yes, it is determined that the matching point pair belongs to the corresponding matching lattice pair; if not, it is determined that the matching point pair does not belong to the corresponding matching lattice pair.
The identification device according to any one of claims 6 to 8, wherein the second extraction unit comprises:

An adjustment module, configured to adjust a sampling threshold according to each of the sub-regions, where the sampling threshold is used to determine the target feature point;

And a fourth determining module, configured to determine a target point in each of the sub-regions as a target feature point, where a parameter value of the target point is greater than the sampling threshold.
The identification device according to any one of claims 6 to 8, wherein the preset algorithm is any one of a direction descriptor ORB algorithm, a scale invariant feature conversion SIFT algorithm or a fast robust feature SURF algorithm. .
An apparatus for identifying a palm vein, comprising: a memory and at least one processor, wherein the memory stores an instruction, and the memory and the at least one processor are interconnected by a line;

The at least one processor invokes the instructions in the memory to cause the palm vein identification device to perform the method of any of claims 1-5.
A computer apparatus, comprising: a processor, the processor for performing the steps of the method of any of claims 1-5 when executing a computer program stored in a memory.
A computer readable storage medium having stored thereon a computer program, wherein the computer program is executed by a processor to perform the steps of the method of any of claims 1-5.