A method, apparatus, terminal and storage medium for fingerprint recognition

Abstract

Embodiments of the present invention disclose a method, apparatus, terminal, and storage medium for fingerprint recognition. The method includes: obtaining an original fingerprint image; performing image enhancement processing on the original fingerprint image to generate a first image; determining the endpoints and bifurcation points of fingerprint ridges on the first image; determining inflection points on the fingerprint ridges based on the endpoints and the bifurcation points; performing feature dimensionality reduction on the endpoints, the bifurcation points, and the inflection points on a second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image; performing fingerprint recognition based on the feature vector. In this solution, by adding inflection points of fingerprint ridges together with endpoints and bifurcation points, the fingerprint features are enriched, making the fingerprint features more reasonable, and performing dimensionality reduction on the inflection points, endpoints, and bifurcation points, thereby reducing memory and improving the comparison speed during fingerprint recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of fingerprint recognition, and in particular, to a method, apparatus, terminal and storage medium for fingerprint recognition. Background Art

[0002] Fingerprint recognition mainly involves the process of extracting key feature information in a fingerprint image, and describing and registering and recognizing the feature information.

[0003] Currently, fingerprint recognition, especially small fingerprint recognition, is applied in many fields. In order to improve the accuracy, the amount of information extracted from small area array fingerprint images is usually large, which results in a long time consumption. Specifically, in the current small area array fingerprint registration field, there are mainly registration methods such as extreme points, ridge lines, and minutiae points. The extreme point registration extracts a lot of information and has a relatively high matching accuracy, but it consumes a large amount of time and memory and has certain requirements for the device; the ridge line feature describes the texture feature of the fingerprint, but lacks the gray information of the original fingerprint image; the minutiae point feature is relatively accurate, but the number of minutiae point features on the small area array is small, which affects the final accuracy.

[0004] Therefore, there is a need for a better method to solve the problems in the prior art. Summary of the Invention

[0005] In view of this, the present invention provides a method, apparatus, terminal and storage medium for fingerprint recognition to solve the problems in the prior art.

[0006] Specifically, the present invention provides the following specific embodiments:

[0007] An embodiment of the present invention provides a method for fingerprint recognition, including:

[0008] Obtaining an original fingerprint image;

[0009] Performing image enhancement processing on the original fingerprint image to generate a first image;

[0010] Determining the endpoints and fork points of the fingerprint ridges on the first image;

[0011] Determining the inflection points on the fingerprint ridges based on the endpoints and the fork points;

[0012] Performing feature dimensionality reduction on the endpoints, the fork points and the inflection points on a second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image;

[0013] Performing fingerprint recognition based on the feature vector.

[0014] In a specific embodiment, the image enhancement processing includes one or more of the following: gray-scale stretching processing, gray-scale enhancement processing, directional filtering processing, binarization processing, and image thinning processing.

[0015] In a specific embodiment, determining the inflection point on the fingerprint ridge line based on the endpoint and the fork point includes:

[0016] If the fingerprint ridge line does not contain the fork point and the length of the fingerprint ridge line exceeds a preset length threshold, then set the fingerprint ridge line as the ridge line to be processed;

[0017] If the fingerprint ridge line contains the fork point, then divide the fingerprint ridge line into three segmented ridge lines based on the fork point; if the length of the segmented ridge line exceeds the preset length threshold, then set the segmented ridge line as the ridge line to be processed;

[0018] For each of the ridge lines to be processed, if the angle between the first connection line from the middle point to the first point and the second connection line from the middle point to the second point is within a preset angle range, then determine the middle point as the inflection point; the first point is the point close to one side endpoint on the ridge line to be processed; the second point is the point close to the other side endpoint on the ridge line to be processed; the distances between the first point and the endpoint it is close to and between the second point and the endpoint it is close to are both greater than a preset distance threshold; the distances from the first point and the second point to the middle point are the same.

[0019] In a specific embodiment, performing feature dimensionality reduction on the endpoint, the fork point, and the inflection point on the second image to obtain the dimensionality-reduced feature vector includes:

[0020] Performing alignment processing on the endpoint, the fork point, and the inflection point on the second image;

[0021] Determining the feature descriptor of the region where the aligned endpoint, fork point, and inflection point are located;

[0022] Performing feature dimensionality reduction by multiplying the feature descriptor with a preset projection matrix to obtain the dimensionality-reduced feature vector.

[0023] In a specific embodiment, performing alignment processing on the endpoint, the fork point, and the inflection point on the second image includes:

[0024] Determining the preset directions of the endpoint, the fork point, and the inflection point on the second image; the preset directions of the endpoint, the fork point, and the inflection point are different from each other;

[0025] Rotate the region where the endpoint, the fork point, and the inflection point are located based on the direction, so that the region where the three are located faces the same direction to achieve alignment.

[0026] An embodiment of the present invention also provides a fingerprint recognition device, including:

[0027] An acquisition module for acquiring an original fingerprint image;

[0028] An image enhancement module, which includes performing image enhancement processing on the original fingerprint image to generate a first image;

[0029] A determination module for determining the endpoints and fork points of the fingerprint ridges on the first image;

[0030] An inflection point module for determining the inflection points on the fingerprint ridges based on the endpoints and the fork points;

[0031] A dimensionality reduction module for performing feature dimensionality reduction on the endpoints, the fork points, and the inflection points on a second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image;

[0032] A fingerprint recognition module for performing fingerprint recognition based on the feature vector.

[0033] In a specific embodiment, the image enhancement processing includes one or more of: gray level stretching processing, gray level enhancement processing, directional filtering processing, binarization processing, and image thinning processing.

[0034] In a specific embodiment, the inflection point module is used for:

[0035] If the fingerprint ridge does not include the fork point and the length of the fingerprint ridge exceeds a preset length threshold, then set the fingerprint ridge as a ridge to be processed;

[0036] If the fingerprint ridge includes the fork point, then divide the fingerprint ridge into three divided ridges based on the fork point; if the length of the divided ridge exceeds the preset length threshold, then set the divided ridge as a ridge to be processed;

[0037] For each ridge to be processed, if the included angle between the first connection line from the middle point to the first point and the second connection line from the middle point to the second point is within a preset angle range, then determine the middle point as an inflection point; the first point is a point close to one endpoint on the ridge to be processed; the second point is a point close to the other endpoint on the ridge to be processed; the distances between the first point and the endpoint it is close to and between the second point and the endpoint it is close to are both greater than a preset distance threshold; the distances from the first point and the second point to the middle point are the same.

[0038] An embodiment of the present invention also provides a terminal, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the above fingerprint recognition method is implemented.

[0039] An embodiment of the present invention also provides a storage medium, in which a computer program is stored, and when the computer program is executed, the above fingerprint recognition method is implemented.

[0040] Therefore, an embodiment of the present invention provides a fingerprint recognition method, device, terminal and storage medium. The method includes: obtaining an original fingerprint image; performing image enhancement processing on the original fingerprint image to generate a first image; determining the endpoints and fork points of fingerprint ridges on the first image; determining the inflection points on the fingerprint ridges based on the endpoints and the fork points; performing feature dimensionality reduction on the endpoints, the fork points and the inflection points on a second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image; performing fingerprint recognition based on the feature vector. In this solution, by adding the inflection points of the fingerprint ridges together with the endpoints and fork points, the fingerprint features are enriched, making the fingerprint features more reasonable, and performing dimensionality reduction on the inflection points, endpoints and fork points, so as to reduce memory and improve the comparison speed during fingerprint recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the protection scope of the present invention. In each drawing, similar components are numbered similarly.

[0042] Figure 1 Shows a schematic flow chart of a fingerprint recognition method proposed by an embodiment of the present invention;

[0043] Figure 2 Shows a schematic diagram of the endpoints and fork points of a fingerprint image in a fingerprint recognition method proposed by an embodiment of the present invention;

[0044] Figure 3 Shows a schematic diagram of determining inflection points in a fingerprint recognition method proposed by an embodiment of the present invention;

[0045] Figure 4 Shows a schematic structural diagram of a fingerprint recognition device proposed by an embodiment of the present invention.

[0046] Legend Explanation:

[0047] 1 - Endpoint; 2 - Fork point;

[0048] 201 - Acquisition module; 202 - Image enhancement module; 203 - Determination module; 204 - Inflection point module;

[0049] 205 - Dimensionality reduction module; 206 - Fingerprint recognition module. Detailed implementation mode

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0051] Generally, the components of the embodiments of the present invention described and illustrated herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but only represents the selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0052] Hereinafter, the terms "including", "having" and their cognates that can be used in various embodiments of the present invention are only intended to represent specific features, numbers, steps, operations, elements, components or combinations of the foregoing items, and should not be construed as first excluding the existence of one or more other features, numbers, steps, operations, elements, components or combinations of the foregoing items or increasing the possibility of one or more features, numbers, steps, operations, elements, components or combinations of the foregoing items.

[0053] In addition, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0054] Unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which various embodiments of the present invention belong. The terms (such as those defined in a general use dictionary) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or being overly formal, unless clearly defined in various embodiments of the present invention.

[0055] Embodiment 1

[0056] Embodiment 1 of the present invention discloses a method for fingerprint recognition, as Figure 1 shown, including the following steps:

[0057] Step S101, acquire the original fingerprint image;

[0058] Specifically, the original fingerprint image is the image of the fingerprint obtained by the fingerprint input device.

[0059] After obtaining the original fingerprint image, perform smoothing denoising on the original fingerprint image to obtain a second image, and execute step S102 to obtain a first image.

[0060] Step S102: Perform image enhancement processing on the original fingerprint image to generate a first image;

[0061] Specifically, the image enhancement processing includes one or more of gray-scale stretching processing, gray-scale enhancement processing, directional filtering processing, binarization processing, and image thinning processing.

[0062] Specifically, the gray-scale stretching processing is used to stretch the gray scale of the original fingerprint image to 0-255 to expand the pixel-level range; the gray-scale enhancement processing is used to enhance the original fingerprint image according to the difference between the original fingerprint image and the smoothed image (the smoothed image is obtained by smoothing the original fingerprint image); the directional filtering processing first calculates the pixel direction values within a certain size square (8*8) point by point to obtain a direction map, and then traverses each pixel, first performing smoothing processing in the direction of the current pixel and then sharpening processing in the normal direction.

[0063] Step S103: Determine the endpoints and fork points of the fingerprint ridges on the first image;

[0064] Specifically, the fingerprint ridges are as Figure 2 shown, where the two ends of the fingerprint ridges are the endpoints 1 of endpoint 1. If there is a crossing point, that is, fork point 2; according to actual experience, the fingerprint ridges in the fingerprint image will only have a three-pronged crossing point, that is, Figure 2 fork point 2 in, and there will be no crossing points with other numbers.

[0065] Step S104: Determine the inflection points on the fingerprint ridges based on the endpoints and the fork points;

[0066] Specifically, as Figure 3 shown, the determining of the inflection points on the fingerprint ridges based on the endpoints and the fork points in step S104 includes:

[0067] If the fingerprint ridge does not contain the fork point and the length of the fingerprint ridge exceeds a preset length threshold, then set the fingerprint ridge as a ridge line to be processed;

[0068] If the fingerprint ridge contains the fork point, then divide the fingerprint ridge into three segmented ridge lines based on the fork point; if the length of the segmented ridge line exceeds the preset length threshold, then set the segmented ridge line as a ridge line to be processed;

[0069] For each of the to-be-processed ridge lines, if the angle between the first connection line from the intermediate point to the first point and the second connection line from the intermediate point to the second point is within a preset angle range, then the intermediate point is determined as an inflection point; the first point is the point close to one end point on the to-be-processed ridge line; the second point is the point close to the other end point on the to-be-processed ridge line; the distances between the first point and the end point it is close to (specifically, the distance on the ridge line) and between the second point and the end point it is close to (specifically, the distance on the ridge line) are both greater than a preset distance threshold; the distances from the first point and the second point to the intermediate point are the same.

[0070] Specifically, start tracking the ridge line from the end point, extract the inflection point on the ridge line longer than a predetermined length (d = 17 - 20) pixels. Assume B is the current tracking point, pixel point A with a length of d1 (for example, 6 - 7) pixels in front of B and pixel point C with a length of d2 (for example, 6 - 7) pixels behind B. Calculate the change angle before and after this point, that is, the angle between AB and AC. If the angle change is too small, it is not an inflection point. If the angle change reaches a certain value, then point B is determined as an inflection point; finally, filter out the inflection points with close distances, low quality, and smaller angle changes. Specifically, the angle threshold for determining whether the angle is large or small can be set according to the actual situation.

[0071] Step S105: Perform feature dimensionality reduction on the end points, fork points, and inflection points on the second image to obtain the dimensionality-reduced feature vectors; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image.

[0072] Specifically, the performing feature dimensionality reduction on the end points, fork points, and inflection points on the second image in step S105 to obtain the dimensionality-reduced feature vectors includes:

[0073] Perform alignment processing on the end points, fork points, and inflection points on the second image;

[0074] Determine the feature descriptors of the region where the aligned end points, fork points, and inflection points are located;

[0075] Perform feature dimensionality reduction by multiplying the feature descriptors with a preset projection matrix to obtain the dimensionality-reduced feature vectors.

[0076] Among them, performing alignment processing on the end points, fork points, and inflection points on the second image includes:

[0077] Determine the preset directions of the end points, fork points, and inflection points on the second image; the preset directions of the end points, fork points, and inflection points are different from each other;

[0078] Rotate the region where the endpoint, the fork point, and the inflection point are located based on the direction, so that the region where the three are located faces the same direction to achieve alignment.

[0079] Specifically, first determine the directions of the endpoint, the fork point, and the inflection point. The direction of the endpoint is the direction from the endpoint to the point at a distance d on the ridge line from the current position; the direction of the fork point is: the direction from the fork point position to the midpoint position of the two ridge lines with a smaller angle; the direction of the inflection point is: the angular bisector direction of the angle between BA and BC (there are two such angles, and here it is the smaller one).

[0080] After determining the directions, the endpoint, the fork point, and the inflection point are all set as feature points. Rotate the feature points (including 15*15 pixels in the surrounding block) to the same direction, calculate the feature descriptor of this region, and then multiply by the projection matrix to reduce the dimension to obtain the reduced-dimension feature.

[0081] Specifically, the projection matrix is calculated in advance through a large amount of data (feature descriptors). For example, the calculation method is to calculate 15,000 feature point descriptors as training samples to form the original feature matrix 15,000*450, calculate the covariance matrix N of the matrix, calculate the eigenvectors of the covariance matrix N, sort according to the magnitude of the eigenvalues, and select the first n eigenvectors (n <= 450, the higher the more the energy percentage of the eigenvalue), to form the projection matrix T.

[0082] Step S106, perform fingerprint recognition based on the eigenvectors.

[0083] This solution requires a small amount of inflection point feature quantities to be determined, has a good effect, and is a position with prominent features in the ridge line. Compared with the extreme points, it has higher accuracy and fewer quantities. Compared with the ridge line features that need to preserve all ridge line information, this method only extracts important features, reduces memory and redundancy. The size of the reduced-dimension feature descriptor is between 10 and 22. Compared with the original size, the memory is reduced by 20 to 30 times, and it has separability through distance comparison, with a good effect, reducing memory and improving the comparison speed.

[0084] Embodiment 2

[0085] To further illustrate this solution, Embodiment 2 of the present invention also discloses a fingerprint recognition device, as Figure 4 shown, including:

[0086] An acquisition module 201, configured to acquire an original fingerprint image;

[0087] An image enhancement module 202, including performing image enhancement processing on the original fingerprint image to generate a first image;

[0088] A determination module 203, configured to determine the endpoints and fork points of the fingerprint ridge lines on the first image;

[0089] An inflection point module 204, configured to determine an inflection point on the fingerprint ridge line based on the endpoint and the fork point;

[0090] A dimensionality reduction module 205, configured to perform feature dimensionality reduction on the endpoint, the fork point, and the inflection point on the second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image;

[0091] A fingerprint recognition module 206, configured to perform fingerprint recognition based on the feature vector.

[0092] In a specific embodiment, the image enhancement processing includes one or more of: gray level stretching processing, gray level enhancement processing, directional filtering processing, binarization processing, and image thinning processing.

[0093] In a specific embodiment, the inflection point module 204 is configured to:

[0094] If the fingerprint ridge line does not include the fork point and the length of the fingerprint ridge line exceeds a preset length threshold, then set the fingerprint ridge line as a ridge line to be processed;

[0095] If the fingerprint ridge line includes the fork point, then divide the fingerprint ridge line into three divided ridge lines based on the fork point; if the length of the divided ridge line exceeds a preset length threshold, then set the divided ridge line as a ridge line to be processed;

[0096] For each of the ridge lines to be processed, if the included angle between a first connection line from an intermediate point to a first point and a second connection line from the intermediate point to a second point is within a preset angle range, then determine the intermediate point as an inflection point; the first point is a point close to one side endpoint on the ridge line to be processed; the second point is a point close to the other side endpoint on the ridge line to be processed; the distances between the first point and the endpoint it is close to and between the second point and the endpoint it is close to are both greater than a preset distance threshold; the distances from the first point and the second point to the intermediate point are the same.

[0097] In a specific embodiment, the dimensionality reduction module 205 is configured to:

[0098] Perform alignment processing on the endpoint, the fork point, and the inflection point on the second image;

[0099] Determine a feature descriptor of the region where the aligned endpoint, fork point, and inflection point are located;

[0100] Perform feature dimensionality reduction by multiplying the feature descriptor by a preset projection matrix to obtain a dimensionality-reduced feature vector.

[0101] In a specific embodiment, the dimensionality reduction module 205 performs alignment processing on the endpoints, fork points, and inflection points on the second image, including:

[0102] Determine the preset directions of the endpoints, fork points, and inflection points on the second image; the preset directions of the endpoints, fork points, and inflection points are different from each other;

[0103] Rotate the regions where the endpoints, fork points, and inflection points are located based on the directions, so that the regions where the three are located face the same direction to achieve alignment.

[0104] Embodiment 3

[0105] Embodiment 3 of the present invention also discloses a terminal, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the fingerprint recognition method described in Embodiment 1 is implemented.

[0106] Specifically, the terminal may be a fingerprint recognition device.

[0107] Embodiment 4

[0108] Embodiment 4 of the present invention also discloses a storage medium. A computer program is stored in the storage medium, and when the computer program is executed, the fingerprint recognition method described in Embodiment 1 is implemented.

[0109] Thus, embodiments of the present invention propose a fingerprint recognition method, device, terminal, and storage medium. The method includes: obtaining an original fingerprint image; performing image enhancement processing on the original fingerprint image to generate a first image; determining the endpoints and fork points of the fingerprint ridges on the first image; determining the inflection points on the fingerprint ridges based on the endpoints and fork points; performing feature dimensionality reduction on the endpoints, fork points, and inflection points on a second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image; performing fingerprint recognition based on the feature vector. In this solution, by adding inflection points of the fingerprint ridges together with the endpoints and fork points, the fingerprint features are enriched, making the fingerprint features more reasonable, and performing dimensionality reduction on the inflection points, endpoints, and fork points, thereby reducing memory and improving the comparison speed during fingerprint recognition.

[0110] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and structure diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in an alternative implementation, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the structure diagram and / or flowchart, as well as the combination of blocks in the structure diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0111] In addition, each functional module or unit in various embodiments of the present invention can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0112] If the described function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0113] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention.

Claims

1. A method for fingerprint recognition, characterized in that, comprising: obtaining an original fingerprint image; performing image enhancement processing on the original fingerprint image to generate a first image; determining endpoints and fork points of fingerprint ridges on the first image; determining inflection points on the fingerprint ridges based on the endpoints and the fork points; wherein, if the fingerprint ridge does not contain the fork point and the length of the fingerprint ridge exceeds a preset length threshold, then setting the fingerprint ridge as a ridge line to be processed; if the fingerprint ridge contains the fork point, then dividing the fingerprint ridge into three divided ridge lines based on the fork point; if the length of the divided ridge line exceeds the preset length threshold, then setting the divided ridge line as a ridge line to be processed; for each of the ridge lines to be processed, if the angle between a first connection line from a middle point to a first point and a second connection line from the middle point to a second point is within a preset angle range, then determining the middle point as an inflection point; the first point is a point close to one endpoint on the ridge line to be processed; the second point is a point close to the other endpoint on the ridge line to be processed; the distances between the first point and the endpoint it is close to and between the second point and the endpoint it is close to are both greater than a preset distance threshold; the distances from the first point and the second point to the middle point are the same; performing feature dimensionality reduction on the endpoints, the fork points and the inflection points on a second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image; performing fingerprint recognition based on the feature vector.

2. The method according to claim 1, characterized in that, the image enhancement processing includes one or more of gray level stretching processing, gray level enhancement processing, directional filtering processing, binarization processing and image thinning processing.

3. The method according to claim 1, characterized in that, the performing feature dimensionality reduction on the endpoints, the fork points and the inflection points on the second image to obtain a dimensionality-reduced feature vector includes: performing alignment processing on the endpoints, the fork points and the inflection points on the second image; determining a feature descriptor of the region where the aligned endpoints, fork points and inflection points are located; performing feature dimensionality reduction by multiplying the feature descriptor with a preset projection matrix to obtain a dimensionality-reduced feature vector.

4. The method according to claim 1, characterized in that, the performing alignment processing on the endpoints, the fork points and the inflection points on the second image includes: determining preset directions of the endpoints, the fork points and the inflection points on the second image; the preset directions of the endpoints, the fork points and the inflection points are different from each other; rotating the regions where the endpoints, the fork points and the inflection points are located based on the directions so that the regions face the same direction to achieve alignment.

5. A fingerprint recognition device, characterized in that, comprising: an obtaining module for obtaining an original fingerprint image; an image enhancement module for performing image enhancement processing on the original fingerprint image to generate a first image; a determining module for determining endpoints and fork points of fingerprint ridges on the first image; An inflection point module, configured to determine an inflection point on the fingerprint ridge line based on the endpoint and the fork point; The inflection point module is further configured to: If the fingerprint ridge line does not include the fork point and the length of the fingerprint ridge line exceeds a preset length threshold, set the fingerprint ridge line as a ridge line to be processed; If the fingerprint ridge line includes the fork point, divide the fingerprint ridge line into three divided ridge lines based on the fork point; if the length of the divided ridge line exceeds the preset length threshold, set the divided ridge line as a ridge line to be processed; For each of the ridge lines to be processed, if the angle between the first connection line from the middle point to the first point and the second connection line from the middle point to the second point is within a preset angle range, determine the middle point as an inflection point; the first point is a point close to one endpoint on the ridge line to be processed; the second point is a point close to the other endpoint on the ridge line to be processed; the distances between the first point and the endpoint it is close to and between the second point and the endpoint it is close to are both greater than a preset distance threshold; The distances from the first point and the second point to the middle point are the same; A dimensionality reduction module, configured to perform feature dimensionality reduction on the endpoint, the fork point, and the inflection point on the second image to obtain a dimensionality-reduced feature vector; the second image is obtained by performing smoothing and denoising processing on the original fingerprint image; A fingerprint recognition module, configured to perform fingerprint recognition based on the feature vector.

6. The device according to claim 5, wherein the image enhancement processing includes one or more of gray level stretching processing, gray level enhancement processing, directional filtering processing, binarization processing, and image thinning processing.

7. A terminal, wherein it includes a memory and a processor, a computer program is stored in the memory, and when the processor executes the computer program, the fingerprint recognition method according to any one of claims 1-4 is implemented.

8. A storage medium, wherein a computer program is stored in the storage medium, and when the computer program is executed, the fingerprint recognition method according to any one of claims 1-4 is implemented.

Images