Optical fingerprint recognition method and system

Abstract

The present invention relates to the technical field of fingerprint recognition and provides an optical fingerprint recognition method and system, the method comprising: obtaining a fingerprint image of a target finger from an optical fingerprint recognition system; performing preliminary fingerprint detail extraction based on the fingerprint image to generate preliminary fingerprint detail information; screening out fingerprint details in the preliminary fingerprint detail information that are located in the edge area of ​​the fingerprint image and have suspicious morphology to generate suspicious fingerprint detail information; obtaining image information of neighboring fingers of the target finger; analyzing the interference cause correlation between the suspicious fingerprint detail information and the neighboring finger image information to generate corresponding interference cause correlation analysis results; and correcting the suspicious fingerprint detail information during the fingerprint recognition process based on the interference cause correlation analysis results to correct fingerprint feature extraction errors caused by interference from neighboring fingers. The present invention has the effect of improving the recognition accuracy and reliability of the optical fingerprint recognition system in multi-finger input scenarios and reducing the risk of misjudgment.

Description

Technical Field

[0001] The present invention relates to the technical field of fingerprint recognition, and in particular to an optical fingerprint recognition method and system. Background Art

[0002] In situations where rapid and continuous verification of personnel identities is required, such as at employee attendance points in large factories or at entrance gates to convention and exhibition centers, optical fingerprint recognition devices that support simultaneous multi-finger input are often deployed to improve access efficiency. After the illumination system is activated, the photosensitive element synchronously captures raw fingerprint images of all pressed fingers. This image data is then fed into a parallel image processing design consisting of multiple image processing pathways, each of which is responsible for independently processing the raw fingerprint images of one or more assigned fingers. This includes image quality assessment, preprocessing, and feature point extraction and encoding, ultimately generating digitized fingerprint feature data.

[0003] In an optical fingerprint recognition system that uses a large-area optical photosensitive plate and supports simultaneous multi-finger input, when a user presses multiple fingers on the photosensitive plate simultaneously with a small spacing, the mutual interference of light from adjacent fingers on the imaging optical path of the target finger and the transmission effect of local micro-deformation of the photosensitive plate caused by the simultaneous pressing of multiple fingers may cause non-intrinsic artifact signals or local geometric distortion to be superimposed on the original fingerprint image of the target finger. This may cause the extracted fingerprint feature template to contain false feature points, omit real feature points, or cause the feature point parameters to be globally inaccurate. Ultimately, how to avoid this problem of reduced overall fingerprint recognition system accuracy and increased risk of misjudgment caused by improper handling of optical and physical interactions between fingers by parallel processing pathways?

[0004] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0005] The purpose of this application is to provide an optical fingerprint recognition method and system, which has the advantages of improving the recognition accuracy and reliability of the optical fingerprint recognition system in multi-finger input scenarios and reducing the risk of misjudgment.

[0006] This application provides an optical fingerprint recognition method, the technical solution is as follows:

[0007] include:

[0008] acquiring a fingerprint image of a target finger from an optical fingerprint recognition system;

[0009] Perform preliminary fingerprint detail extraction based on the fingerprint image to generate preliminary fingerprint detail information;

[0010] Screening out the fingerprint details that are located in the edge area of ​​the fingerprint image and have suspicious morphology in the preliminary fingerprint detail information to generate suspicious fingerprint detail information;

[0011] Obtaining image information of fingers adjacent to the target finger;

[0012] Analyze the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information, and generate corresponding interference cause correlation analysis results;

[0013] Based on the results of the correlation analysis of interference causes, the suspicious fingerprint details in the fingerprint recognition process are corrected to correct the fingerprint feature extraction errors caused by interference from neighboring fingers.

[0014] Through the above solution, the recognition accuracy and reliability of the optical fingerprint recognition system in multi-finger input scenarios are improved, and the risk of misjudgment is reduced.

[0015] Furthermore, the present application also proposes that the steps of analyzing the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information and generating the corresponding interference cause correlation analysis results include:

[0016] Identify fingerprint detail clustering areas corresponding to suspicious fingerprint detail information in edge areas of the fingerprint image;

[0017] Extract the tissue morphological features of the fingerprint minutiae clustering area;

[0018] Based on the image information of the neighboring fingers, obtaining the position information of the neighboring fingers;

[0019] Based on the position information, determine whether the fingerprint detail gathering area is within the potential optical influence range of the adjacent finger and generate a position judgment result;

[0020] Compare the tissue morphological features with the preset pseudo-feature cluster morphological rules to generate morphological comparison results;

[0021] A correlation analysis is performed based on the morphological comparison result and the position judgment result. When the morphological comparison result and the position judgment result both indicate correlation with interference from adjacent fingers, it is determined that there is a correlation, and an interference cause correlation analysis result is generated based on the correlation analysis result.

[0022] Through the above scheme, a more specific interference cause correlation analysis method is provided, which improves the accuracy of the analysis.

[0023] Furthermore, the present application also proposes to screen out fingerprint details with suspicious morphology located in the edge area of ​​the fingerprint image from the preliminary fingerprint detail information, and the steps of generating the suspicious fingerprint detail information include:

[0024] Obtaining a fingerprint area contour line of the fingerprint image, and expanding inward by a preset pixel width based on the fingerprint area contour line to determine an edge area of ​​the fingerprint image;

[0025] In the edge area, local image quality scoring and geometric morphology analysis are performed on each fingerprint detail to generate a corresponding suspicion score;

[0026] The corresponding fingerprint details whose suspicion score exceeds the preset threshold are sorted into suspicious fingerprint detail information.

[0027] Through the above scheme, a more specific method for screening suspicious fingerprint details is provided, thereby improving the effectiveness of screening.

[0028] Furthermore, the present application also proposes that the steps of comparing the tissue morphological features with the preset pseudo-feature cluster morphological rules to generate the morphological comparison results include:

[0029] Obtaining external environmental condition information corresponding to when the optical fingerprint recognition system collects fingerprint images and characteristic information of attachments corresponding to adjacent fingers;

[0030] Determining whether a preset pseudo-feature cluster morphology rule is suitable for the pseudo-feature cluster morphology caused by the concealed optical effect of the adjacent finger under conditions corresponding to the external environmental condition information and the attachment characteristic information, and generating a rule adaptability judgment result;

[0031] If the rule adaptability judgment result is negative, the pseudo-feature cluster morphology rule is adjusted and updated according to the external environmental condition information and the attachment characteristic information to generate the current comparison rule;

[0032] If the rule adaptability judgment result is yes, the pseudo-feature cluster morphology rule is directly used as the current comparison rule;

[0033] Compare the tissue morphological features with the current comparison rules to generate morphological comparison results.

[0034] Through the above scheme, a method for dynamically adjusting the comparison rules according to the external environment and attachment information is provided, thereby improving the robustness of the morphological comparison.

[0035] Furthermore, the present application also proposes that, based on the external environmental condition information and the attachment characteristic information, the pseudo-feature cluster morphology rules are adjusted and updated, and the steps of generating the current comparison rules include:

[0036] Based on the attachment characteristic information, determining whether at least two types of attachments coexist on the surface of the adjacent finger, or whether characteristic parameters of a single type of attachment on the surface of the adjacent finger change over time, and generating an attachment status determination result;

[0037] If the result of the attachment state determination is yes, then for each attachment type in the case where at least two types of attachments coexist, or for the current characteristics of a single type of attachment in the case where a single type of attachment exists, determining a corresponding basic impact mode based on the attachment characteristic information;

[0038] All basic impact patterns are combined according to the preset superposition logic to form composite impact pattern information.

[0039] According to the external environmental condition information and the composite impact pattern information, the parameters of the preset pseudo-feature cluster morphology rule are adjusted, or the structure of the preset pseudo-feature cluster morphology rule is adjusted to generate the current comparison rule.

[0040] Through the above scheme, a method for handling complex attachment situations and forming composite impact pattern information is provided, making rule adjustments more targeted.

[0041] Furthermore, the present application also proposes that the steps of combining all basic impact patterns according to a preset superposition logic to form composite impact pattern information include:

[0042] Get the preset overlay logic;

[0043] Obtain coupling indication information between all basic impact modes;

[0044] Selecting a coupling effect adjustment rule corresponding to the coupling indication information from a preset coupling effect adjustment rule set;

[0045] Apply coupling effect adjustment rules to adjust the superposition logic;

[0046] Using the adjusted superposition logic, all basic impact patterns are combined to form composite impact pattern information.

[0047] Through the above scheme, a combined method considering the coupling effect between basic impact modes is provided, making the composite impact mode information more accurate.

[0048] Furthermore, the present application also proposes that the step of obtaining coupling indication information between all basic impact modes includes:

[0049] Obtain the expected combined effect characteristics of all basic impact modes determined under the preset benchmark coupling conditions;

[0050] By analyzing the interactive characteristics of basic influence patterns, we can obtain the actual combined effect characteristics when basic influence patterns act together;

[0051] Compare the actual combination effect characteristics with the expected combination effect characteristics to determine effect deviations;

[0052] Determining whether the effect deviation exceeds a preset deviation threshold, and whether the pattern of the effect deviation deviates from any pattern in a preset set of known coupling patterns;

[0053] If the judgment result is yes, coupling indication information is generated based on the deviation condition and the offset condition.

[0054] Through the above scheme, a method is provided to identify whether there are unexpected coupling effects between basic impact modes, providing a basis for rule adjustments.

[0055] Furthermore, the present application also proposes that the step of determining whether the mode of the effect deviation deviates from any mode in a preset set of known coupling modes includes:

[0056] The pattern of effect deviation is analyzed by components, and the components that are consistent with the known coupling pattern set are taken as the first effect component, and the components that are different from the known coupling pattern set are taken as the second effect component;

[0057] Obtain preset component impact quantification rules;

[0058] Apply the component influence quantification rule to determine the influence degree corresponding to the first effect component and the influence degree corresponding to the second effect component;

[0059] According to the preset attribution determination logic, based on the influence degree of the first effect component and the influence degree of the second effect component, it is determined whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set.

[0060] Through the above scheme, a method is provided to quantify the effect deviation component and determine whether its pattern deviates from the known coupling pattern, thereby improving the accuracy of the coupling indication information.

[0061] Furthermore, the present application also proposes that, according to a preset attribution determination logic, based on the influence of the first effect component and the influence of the second effect component, the step of determining whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set includes:

[0062] Obtaining physical characteristics of adjacent fingers and imaging parameters of optical fingerprint recognition systems;

[0063] Determining a judgment parameter based on the physical characteristics and the imaging parameters;

[0064] According to the determination parameter, based on the influence degree of the first effect component and the influence degree of the second effect component, it is determined whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set.

[0065] Through the above solution, a method for attribution determination combining physical characteristics and imaging parameters is provided, making the determination process more accurate.

[0066] An optical fingerprint recognition system for performing optical fingerprint recognition, comprising:

[0067] A fingerprint image acquisition module, used to acquire a fingerprint image of a target finger from an optical fingerprint recognition system;

[0068] A detail information extraction module is used to extract preliminary fingerprint details based on the fingerprint image and generate preliminary fingerprint detail information;

[0069] A suspicious fingerprint screening module is used to screen out the fingerprint details that are located in the edge area of ​​the fingerprint image and have suspicious morphology in the preliminary fingerprint detail information, and generate the suspicious fingerprint detail information;

[0070] A neighboring finger image acquisition module is used to acquire neighboring finger image information of fingers adjacent to the target finger;

[0071] Interference correlation analysis module, used to analyze the interference cause correlation between the suspicious fingerprint details and the adjacent finger image information, and generate corresponding interference cause correlation analysis results;

[0072] The recognition process correction module is used to correct the suspicious fingerprint details during the fingerprint recognition process based on the interference cause correlation analysis results to correct the fingerprint feature extraction errors caused by interference from neighboring fingers.

[0073] Through the above solution, a system for implementing the above method is provided, which is convenient for practical application.

[0074] From the above, it can be seen that the optical fingerprint recognition method and system provided by the present application analyzes the correlation between the interference causes of suspicious fingerprint details and the image information of neighboring fingers, and corrects the suspicious details in the fingerprint recognition process based on this, thereby correcting the feature extraction errors caused by interference from neighboring fingers, thereby improving the recognition accuracy and reliability of the optical fingerprint recognition system in multi-finger input scenarios and reducing the risk of misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 This is a flow chart of an optical fingerprint recognition method according to one embodiment of the present invention;

[0076] Figure 2 This is one of the flow charts of an optical fingerprint recognition method according to another embodiment of the present invention;

[0077] Figure 3 This is a second flow chart of a method for optical fingerprint recognition according to another embodiment of the present invention;

[0078] Figure 4 This is a third flow chart of a method for optical fingerprint recognition according to another embodiment of the present invention;

[0079] Figure 5 This is a fourth flow chart of a method for optical fingerprint recognition according to another embodiment of the present invention;

[0080] Figure 6 This is a fifth flow chart of a method for optical fingerprint recognition according to another embodiment of the present invention;

[0081] Figure 7 This is a sixth flow chart of a method for optical fingerprint recognition according to another embodiment of the present invention;

[0082] Figure 8 This is a seventh flow chart of a method for optical fingerprint recognition according to another embodiment of the present invention;

[0083] Figure 9 This is a flowchart of an optical fingerprint recognition method according to another embodiment of the present invention.

[0084] Figure 10 This is a system block diagram of an optical fingerprint recognition system in another embodiment of the present invention;

[0085] Description of reference numerals:

[0086] 1. Optical fingerprint recognition system; 11. Fingerprint image acquisition module; 12. Detail information extraction module; 13. Suspicious fingerprint screening module; 14. Neighboring finger image acquisition module; 15. Interference correlation analysis module; 16. Recognition process correction module. DETAILED DESCRIPTION

[0087] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0088] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0089] Traditional optical fingerprint recognition systems support simultaneous multi-finger input. When a user presses multiple fingers simultaneously on the photosensitive plate with a small distance between them, the interference between adjacent fingers and the target finger's imaging path, as well as the transmission of local micro-deformations of the photosensitive plate caused by the simultaneous pressing of multiple fingers, can cause non-intrinsic artifact signals or local geometric distortion to be superimposed on the original fingerprint image of the target finger. Under the system's parallel image processing design, when each independent image processing pathway receives a fingerprint image containing such mixed interference between fingers, its standard image preprocessing and feature extraction procedures lack the ability to identify and specifically eliminate this specific interference, potentially misinterpreting it as fingerprint structure or noise. This can result in the extracted fingerprint feature template containing false feature points, missing true feature points, or overall inaccuracy in feature point parameters.

[0090] For example, suppose that in a high-speed traffic scenario, the user needs to quickly press the index finger, middle finger, and ring finger of the same hand on the photosensitive plate of the optical fingerprint recognition device at the same time. Since the user's operation speed is fast and may not strictly follow the standard posture, the distance between the middle finger and the ring finger and the index finger is very small. At this time, the edge area of ​​the ring finger may block or scatter the imaging light path of the middle finger. At the same time, the combined pressure of the three fingers may cause the photosensitive plate to produce a slight deformation near the imaging area of ​​the middle finger. The superposition of these optical and physical effects causes the edge area of ​​the fingerprint image of the middle finger near the ring finger to have non-fingerprint texture spots or slight distortion of the lines. When processing the middle finger image, the existing image processing pathway may mistakenly identify these spots as the end points or bifurcation points of the fingerprint, or the distortion of the lines may cause the position of the real feature points to shift, generating fingerprint feature data containing erroneous information.

[0091] In this regard, this application proposes an optical fingerprint recognition method, combining Figure 1 Shown, including:

[0092] S1, obtaining a fingerprint image of a target finger from an optical fingerprint recognition system;

[0093] S2, performing preliminary fingerprint detail extraction based on the fingerprint image to generate preliminary fingerprint detail information;

[0094] S3, screening out fingerprint details with suspicious morphology located in the edge area of ​​the fingerprint image from the preliminary fingerprint detail information, and generating suspicious fingerprint detail information;

[0095] S4, obtaining image information of fingers adjacent to the target finger;

[0096] S5, analyzing the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information, and generating a corresponding interference cause correlation analysis result;

[0097] S6, based on the interference cause correlation analysis results, correct the suspicious fingerprint details in the fingerprint recognition process to correct the fingerprint feature extraction errors caused by the interference of adjacent fingers.

[0098] Among them, screening out suspicious fingerprint details located in the edge area of ​​the fingerprint image from the preliminary fingerprint detail information and generating suspicious fingerprint detail information means identifying those details located near the edge of the fingerprint image and whose shape, structure or local texture are abnormal or do not conform to the typical fingerprint feature pattern from the preliminary extracted fingerprint details, and aggregating these details. This can be achieved using a variety of image analysis technologies, such as determining the fingerprint area contour based on image segmentation, and then expanding inward or outward to a certain range to define the edge area; within the edge area, local image quality assessment and geometric morphology analysis are performed on each detail, and a suspicion score is calculated based on these analysis results. Details with scores exceeding the preset standard are marked as suspicious. This is mainly to narrow the scope of subsequent interference analysis, focusing on the areas and details most likely to be affected by neighboring fingers, and improving processing efficiency and accuracy.

[0099] Acquiring image information of neighboring fingers of the target finger refers to collecting image data of other fingers pressed on the photosensitive plate at the same time as the target finger through the optical fingerprint recognition system, or related information extracted therefrom. This can be achieved in the same way as acquiring the fingerprint image of the target finger. For example, when acquiring multiple fingers simultaneously, the system will acquire an overall image containing all fingers, and then use image segmentation technology to separate the areas of different fingers and extract the image parts corresponding to the neighboring fingers. This is mainly to provide information such as the existence, location, shape, and possible surface state of the neighboring fingers, as an external reference and basis for analyzing whether suspicious details in the target finger image are caused by interference from neighboring fingers.

[0100] Analyzing the interference-causing correlation between the suspicious fingerprint detail information and the image information of the neighboring fingers and generating corresponding interference-causing correlation analysis results involves assessing whether there is a causal relationship or strong correlation between the appearance of the suspicious fingerprint detail and the presence and status of the neighboring fingers. This can be achieved using a variety of analytical methods, such as comparing the position of the suspicious detail with the relative position of the neighboring fingers to determine whether the suspicious detail is within the optical influence range of the neighboring fingers; analyzing whether the morphological characteristics of the suspicious detail match the known pseudo-feature patterns generated by the interference of the neighboring fingers; and combining environmental conditions and the characteristics of the surface attachments of the neighboring fingers to assess the possibility and manifestation of these factors interfering with the imaging of the target finger through the neighboring fingers. This is primarily to determine whether the suspicious fingerprint detail is indeed caused by the interference of the neighboring fingers, providing a judgment basis for subsequent corrections.

[0101] Correcting suspicious fingerprint details during the fingerprint recognition process based on the results of the interference-causing correlation analysis to correct fingerprint feature extraction errors caused by interference from neighboring fingers involves processing suspicious fingerprint details that are determined to be caused by interference from neighboring fingers based on the correlation determined by the analysis to eliminate or mitigate their impact on the final fingerprint feature extraction results. This can be achieved using a variety of correction strategies. For example, if a suspicious detail is determined to be a pseudo-feature, it is removed from the list of initially extracted fingerprint details; if a suspicious detail is determined to be a real feature but its parameters are affected by interference, its parameters are corrected; and if interference causes line breakage or adhesion, line repair is attempted. This is primarily intended to improve the accuracy of fingerprint feature extraction, avoid mistaking pseudo-features for real features or omit real features, and thus enhance overall fingerprint recognition performance.

[0102] The solution of this application aims to address the problem of interference caused by neighboring fingers in multi-finger optical fingerprint recognition through a series of coordinated steps. First, the system acquires a fingerprint image of the target finger, which serves as the basis for all subsequent processing. Next, preliminary fingerprint minutiae extraction is performed on this image, obtaining preliminary minutiae information including features such as endpoints and bifurcation points. Considering that interference from neighboring fingers often occurs at the edge of the target finger image, and the artifacts or distortion caused by interference often exhibit unusual morphology, the solution further filters out fingerprint minutiae located at the edge of the fingerprint image and with suspicious morphology from the preliminary minutiae information to generate suspicious fingerprint minutiae information, effectively narrowing the scope of analysis. Simultaneously, the system acquires image information of neighboring fingers pressed simultaneously with the target finger. This information provides external reference for the location and morphology of the neighboring fingers. Crucially, the solution then analyzes the correlation between these suspicious fingerprint minutiae and the image information of the neighboring fingers to determine the cause of the interference. This analysis goes beyond simply determining proximity; rather, it delves into the causal relationship between the morphology and location of the suspicious minutiae and the presence, position, and state of the neighboring fingers, thereby determining whether the suspicious minutiae are caused by optical or physical interference from the neighboring fingers. Based on the results of this correlation analysis, the solution corrects suspicious fingerprint details during the fingerprint recognition process. If the analysis indicates that the suspicious details are indeed artifacts caused by interference from neighboring fingers, they are removed or corrected. If the analysis indicates that they are real features but affected by interference, their parameters are modified. By incorporating information from neighboring fingers and performing correlation analysis, the solution can identify and specifically address interference caused by inter-finger interactions, which is difficult for traditional methods to address. This prevents mistaking artifacts for real features or missing real features, thereby correcting fingerprint feature extraction errors and improving ultimate recognition accuracy.

[0103] In some preferred embodiments, the present application is specifically implemented as follows. First, the optical fingerprint recognition system captures a fingerprint image of the target finger. Next, standard image processing algorithms, such as Gabor filtering, binarization, and thinning, are used to extract preliminary fingerprint details from the fingerprint image, such as the location and direction of endpoints and bifurcation points. In order to screen out suspicious details, the precise outline of the fingerprint area can be determined through image segmentation, and then an area extending inward by a certain pixel width from the outline can be defined as an edge area. Within the edge area, for each preliminary extracted fingerprint detail, the contrast and clarity of its local image are calculated, and the geometric morphological features such as the continuity and directional consistency of the surrounding ridges are analyzed. These indicators are combined to calculate a suspicion score. Details with a score higher than a preset threshold are marked as suspicious fingerprint details. At the same time, the system captures an overall image containing the target finger and adjacent fingers, identifies and extracts the image area of ​​the adjacent fingers through a finger segmentation algorithm, and obtains their location information. Subsequently, the correlation between the suspicious fingerprint details and the image information of the adjacent fingers is analyzed. For example, it determines whether the suspicious detail is located in or near the projection area of ​​a neighboring finger; extracts texture features from the area where the suspicious detail is located and compares them with a preset library of artifact patterns caused by neighboring finger interference; and combines the ambient light information at the time of acquisition with information about whether there are attachments such as water stains and oil on the surface of the user's finger to assess the possibility of generating a specific interference pattern. If the analysis results indicate that the suspicious detail is highly correlated with neighboring finger interference, for example, the suspicious detail is located near the neighboring finger and its shape is highly similar to the known artifact pattern, it is judged as a pseudo-feature caused by neighboring finger interference. Finally, based on this judgment result, when generating the final fingerprint feature template, these suspicious details judged to be pseudo-features are removed from the list of initially extracted details, or their parameters are adjusted to correct feature extraction errors caused by neighboring finger interference.

[0104] Optional, combined Figure 2 As shown, S5 analyzes the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information, and generates the corresponding interference cause correlation analysis result, including:

[0105] S51, identifying a fingerprint detail clustering area corresponding to the suspicious fingerprint detail information in an edge area of ​​the fingerprint image;

[0106] S52, extracting the tissue morphological features of the fingerprint minutiae clustering area;

[0107] S53, obtaining position information of the neighboring fingers based on the image information of the neighboring fingers;

[0108] S54, judging whether the fingerprint detail concentration area is within the potential optical influence range of the adjacent finger based on the position information, and generating a position judgment result;

[0109] S55, comparing the tissue morphological features with the preset pseudo-feature cluster morphological rules to generate a morphological comparison result;

[0110] S56, performing correlation analysis based on the morphological comparison result and the position judgment result. When both the morphological comparison result and the position judgment result indicate that they are related to the interference of the adjacent finger, determining that there is correlation, and generating an interference cause correlation analysis result based on the correlation analysis result.

[0111] Among them, the fingerprint detail clustering area refers to the local area formed by multiple suspicious fingerprint details that are close in space in the edge area of ​​the fingerprint image, which can be implemented by image segmentation algorithm, connected domain analysis or clustering algorithm; the tissue morphological characteristics refer to the arrangement, density, direction, curvature and other geometric and texture properties of details such as lines, endpoints, bifurcation points in the fingerprint detail clustering area, which can be implemented by texture analysis algorithm, morphological feature extraction method or learning-based feature descriptor; the position information of neighboring fingers refers to the spatial coordinates, contour or occupied area information of neighboring fingers in the photosensitive plate or fingerprint image, which can be implemented by image segmentation, contour detection or multi-finger image registration technology; the potential optical influence range refers to the area where neighboring fingers may interfere with the imaging of the target finger through optical paths (such as scattering, reflection) or physical deformation, which can be determined based on optical models, physical models or empirical thresholds; the preset pseudo-feature cluster morphological rules refer to the pre-established typical morphological manifestations of pseudo-features that may be caused by interference from neighboring fingers. A collection of patterns, which can be constructed based on a large amount of experimental data, simulation models or expert experience; the morphological comparison result refers to the matching degree or classification judgment obtained by comparing the tissue morphological characteristics of the extracted fingerprint detail cluster area with the preset pseudo-feature cluster morphological rules, which can be achieved by pattern recognition algorithm, classifier or similarity measurement method; the position judgment result refers to the binary or probabilistic judgment of whether the fingerprint detail cluster area is within the potential optical influence range of the neighboring finger, which can be achieved based on spatial position relationship judgment or distance calculation; correlation analysis refers to the process of comprehensively considering the morphological comparison results and position judgment results to determine whether the suspicious fingerprint detail cluster area has a causal relationship with the interference of the neighboring finger, which can be achieved by logical judgment, decision rules or machine learning models; the interference cause correlation analysis result refers to the output of the correlation analysis process, which indicates the degree of correlation or whether there is a correlation between the suspicious fingerprint detail cluster area and the interference of the neighboring finger. It can be a binary mark, a confidence score or a classification label.

[0112] In some preferred embodiments, the specific implementation is as follows. When identifying the fingerprint detail clustering area corresponding to the suspicious fingerprint detail information in the edge area of ​​the fingerprint image, the suspicious fingerprint detail information can be first binarized, and then the connected domain analysis algorithm can be used to identify mutually connected or close pixel sets, and these sets are determined as fingerprint detail clustering areas. When extracting the tissue morphological characteristics of the fingerprint detail clustering area, the area, perimeter, density, and direction histogram of each clustering area can be calculated, or the texture feature vector can be extracted by applying a Gabor filter group. When obtaining the position information of the neighboring fingers based on the neighboring finger image information, the original multi-finger image can be segmented into finger areas, for example, using a threshold segmentation or edge detection algorithm to obtain the contour line or mask of each finger, thereby determining the pixel coordinate range of the neighboring finger in the image. When judging whether the fingerprint detail clustering area is in the potential optical influence range of the neighboring finger based on the position information, the shortest distance from the center of mass of each clustering area to the contour of the neighboring finger can be calculated and compared with a preset distance threshold. If the distance is less than the threshold, it is judged to be in the potential influence range. When comparing tissue morphological features with preset pseudo-feature cluster morphological rules, a rule library containing a variety of typical pseudo-feature morphological patterns can be established, such as patterns describing blurred spots caused by scattered light and line distortion caused by deformation. The extracted tissue morphological features are then matched with the patterns in the rule library for matching degree or classification. When performing correlation analysis based on the morphological comparison results and position judgment results, a logical judgment rule can be used. For example, if the morphological comparison results indicate that the area highly matches the pseudo-feature pattern, and the position judgment results indicate that the area is within the potential influence range of the adjacent finger, then a correlation is determined.

[0113] Optional, combined Figure 3 As shown, S3 filters out the fingerprint details that are located in the edge area of ​​the fingerprint image and have suspicious morphology in the preliminary fingerprint detail information. The steps of generating the suspicious fingerprint detail information include:

[0114] S31, obtaining a fingerprint area contour line of the fingerprint image, and inwardly expanding a preset pixel width based on the fingerprint area contour line to determine an edge area of ​​the fingerprint image;

[0115] S32, in the edge area, performing local image quality scoring and geometric morphology analysis on each fingerprint detail to generate a corresponding suspicion score;

[0116] S33, organizing the corresponding fingerprint details whose suspicion scores exceed a preset threshold into suspicious fingerprint detail information.

[0117] Among them, the fingerprint area contour line refers to the boundary line of the area actually containing the fingerprint texture in the fingerprint image. Specifically, it can be achieved by using an image segmentation algorithm, such as grayscale threshold, edge detection or active contour model, to distinguish the fingerprint area from the background area and extract the outer boundary of the fingerprint area. Its purpose is to accurately define the range of valid information in the fingerprint image; the preset pixel width refers to a predetermined distance value in pixels, which can be set according to factors such as the imaging characteristics of the optical fingerprint recognition system, finger pressing habits and empirical data. Its purpose is to define a specific width range extending inward from the fingerprint area contour line. This range is considered to be an area that is susceptible to edge effects or interference from neighboring fingers; the local image quality score refers to a numerical value for quantitatively evaluating the image quality of a local area surrounding a fingerprint detail in the fingerprint image. Specifically, it can be obtained by calculating indicators such as contrast, clarity, line direction consistency or Gabor filter response of the local area. Its purpose is to reflect the imaging quality and the degree of noise interference in the area where the fingerprint detail is located; geometric morphology Analysis refers to the inspection and measurement of the shape and structure of the fingerprint minutiae themselves, as well as their relationship with the surrounding ridges. Specifically, this can be done by analyzing geometric properties such as whether a fingerprint minutiae is an endpoint or bifurcation point, the length and curvature of the connected ridge segments, and the angle of the bifurcation, and comparing them with the normal fingerprint minutiae morphology. The purpose is to identify pseudo-minutiae or distorted minutiae that have abnormal morphology and do not conform to typical fingerprint minutiae characteristics. The suspicion score is a quantitative indicator of the likelihood that a fingerprint minutiae is a pseudo-minutiae or distorted minutiae, after comprehensively considering the local image quality score and the results of the geometric morphology analysis. Specifically, the local image quality score and the results of the geometric morphology analysis can be integrated into a single score value through weighted summation, decision trees, or machine learning models. The purpose is to provide a unified standard for measuring the suspicion level of each fingerprint minutiae. The preset threshold is a predetermined critical value for the suspicion score, which can be set based on the desired screening stringency, false positive rate, and false negative rate requirements. Its purpose is to serve as a criterion for determining whether a fingerprint minutiae is sufficiently suspicious to be marked and further processed.

[0118] In some preferred embodiments, the Canny edge detection algorithm combined with contour tracing can be used to obtain the fingerprint region outline. A morphological erosion operation with a set erosion kernel size is then performed to simulate the inward expansion of the outline by a preset pixel width, thereby determining the edge region. Within the edge region, for each initially extracted fingerprint minutiae (e.g., an endpoint or bifurcation point), a small neighborhood image block centered on the minutiae is extracted, and the local contrast variance of the image block is calculated as a local image quality score. Simultaneously, the ridge segments connecting the minutiae are analyzed, for example, by measuring the length and curvature of the ridge segments, or, for bifurcation points, the bifurcation angle. This is then compared with the statistical characteristics of normal fingerprint ridges to determine an index of geometric abnormality. The local image quality score and the geometric abnormality index can be weighted and summed, or input into a pre-trained classifier (e.g., a support vector machine or neural network) to output a suspicion score between 0 and 1. A preset threshold, such as 0.7, is set, and fingerprint minutiae with a suspicion score above 0.7 are marked as suspicious. Their information (location, type, orientation, etc.) is collected to form a list of suspicious fingerprint minutiae.

[0119] Optional, combined Figure 4 As shown, S55 compares the tissue morphological features with the preset pseudo-feature cluster morphological rules to generate the morphological comparison results, including:

[0120] S551, obtaining external environmental condition information corresponding to when the optical fingerprint recognition system collects the fingerprint image and characteristic information of the attachment corresponding to the adjacent finger;

[0121] S552, determining whether a preset pseudo-feature cluster morphology rule is suitable for the pseudo-feature cluster morphology caused by the concealed optical effect of the adjacent finger under conditions corresponding to the external environmental condition information and the attachment characteristic information, and generating a rule adaptability determination result;

[0122] S553, if the result of the rule adaptability judgment is no, then according to the external environmental condition information and the attachment characteristic information, the pseudo-feature cluster morphology rule is adjusted and updated to generate the current comparison rule;

[0123] S554, if the rule adaptability judgment result is yes, then directly use the pseudo-feature cluster morphology rule as the current comparison rule;

[0124] S555, comparing the tissue morphological features with the current comparison rules to generate a morphological comparison result.

[0125] External environmental condition information refers to the environmental factors during fingerprint image acquisition, such as ambient temperature, humidity, and light intensity. This information can be acquired using devices such as temperature sensors, humidity sensors, and light sensors. Attachment characteristic information refers to the properties of attachments present on the surface of adjacent fingers, such as their type (sweat, grease, water droplets, dust, etc.), distribution, thickness, optical reflectivity, and transmittance. This information can be acquired using a spectrum analyzer, a capacitive sensor, or by analyzing images of adjacent fingers. Pseudo-feature cluster morphological rules describe the typical morphological patterns of pseudo-features formed in fingerprint images due to interference from adjacent fingers, such as their shape, size, density, distribution range, and texture characteristics. These rules can be represented using a preset mathematical model, a statistical model, or a model trained based on machine learning. Covert optical effects refer to the optical influences on the imaging optical path of the target finger caused by adjacent fingers and their surface attachments that are difficult to directly predict or observe. These effects include light scattering, reflection, refraction, absorption, or changes in the suppressed state of total internal reflection. These effects can result in unexpected light spots, shadows, or local contrast anomalies in the target finger image. Rule compatibility refers to the degree of match between the pre-defined pseudo-feature cluster morphology rules and the actual pseudo-feature cluster morphology that may occur under specific external environmental conditions and attachment characteristics. This can be expressed using similarity calculations, pattern matching scores, or classifier outputs. The current comparison rule refers to the rule actually used when comparing tissue morphology features with the pseudo-feature cluster morphology rules. This rule may be the original pre-defined rule or an updated rule adjusted based on external environmental conditions and attachment characteristics.

[0126] In some preferred embodiments, specifically, obtaining external environmental condition information can be achieved through an environmental sensor array connected to the optical fingerprint recognition system, which can include a temperature sensor, a humidity sensor, and an ambient light sensor. Attachment characteristic information can be obtained by performing spectral analysis or texture analysis on an image of a neighboring finger, for example, analyzing the spectral reflectance of a specific band in the image or the roughness of the local texture to determine whether sweat or grease is present. Determining whether the preset pseudo-feature cluster morphology rules are suitable can be achieved through a pre-trained classification model that takes environmental condition information and attachment characteristic information as input and outputs a suitability score. If the suitability score is lower than a preset threshold, it is determined to be unsuitable. Adjusting and updating the pseudo-feature cluster morphology rules can be achieved through a rule generation module that selects the most matching rule template from a rule library based on the environmental condition information and attachment characteristic information, or adjusts the parameters in the current rule template, such as adjusting the expected size range or contrast threshold of the pseudo-feature. Comparing the tissue morphological features with the current comparison rules can be achieved through a deep learning-based image matching algorithm, which compares the extracted tissue morphological feature image with the pseudo-feature template represented by the current comparison rule and outputs a similarity score as the morphological comparison result.

[0127] Optional, combined Figure 5 As shown, the step of adjusting and updating the pseudo-feature cluster morphology rule according to the external environment condition information and the attachment characteristic information to generate the current comparison rule in step S553 includes:

[0128] S5531, based on the attachment characteristic information, determining whether at least two types of attachments coexist on the surface of the adjacent finger, or whether characteristic parameters of a single type of attachment on the surface of the adjacent finger change over time, and generating an attachment status determination result;

[0129] S5532: If the result of the attachment status determination is yes, then for each attachment type in the case where at least two types of attachments coexist, or for the current characteristics of a single attachment in the case where there is only one type of attachment, determine a corresponding basic impact mode based on the attachment characteristic information;

[0130] S5533, combining all basic impact patterns according to a preset superposition logic to form composite impact pattern information;

[0131] S5534, adjusting the parameters of the preset pseudo-feature cluster morphology rule or adjusting the structure of the preset pseudo-feature cluster morphology rule according to the external environmental condition information and the composite impact pattern information, to generate the current comparison rule.

[0132] Among them, attachment characteristic information refers to data describing the type, distribution, thickness, optical properties and other characteristics of attachments on the surface of adjacent fingers, which can be obtained by image analysis, spectral analysis or sensor measurement, and its purpose is to quantify the impact of attachments on light propagation; basic impact pattern refers to the typical impact model of attachments of a specific type or state on optical imaging under specific external environmental conditions, which can be represented by a lookup table, mathematical function or pre-trained model, and its purpose is to establish a mapping relationship between attachment characteristics and optical impact; superposition logic refers to the rules or algorithms used to describe how multiple basic impact patterns interact and superimpose to form an overall impact, which can be implemented by linear superposition, nonlinear combination or simulation calculation based on physical models, and its purpose is to simulate the comprehensive optical effects under complex attachment conditions; composite impact pattern information refers to the description of the combination of all basic impact patterns according to superposition logic. The data on the overall impact of neighboring fingers on optical imaging under current conditions can be expressed in the form of impact intensity maps, distortion fields or changes in optical transfer functions, with the aim of comprehensively reflecting the optical interference characteristics under the current environment and attachment status; the parameters of the pseudo-feature cluster morphological rules refer to the numerical values ​​or thresholds used to define the morphological characteristics of the pseudo-feature cluster, such as the weight of the shape descriptor, the range of the texture feature or the threshold for similarity calculation, which can be implemented by numerical adjustment or parameter optimization algorithms, with the aim of fine-tuning the rules to adapt to the changing impact pattern; the structure of the pseudo-feature cluster morphological rules refers to the mathematical model or algorithm framework used to describe the morphological characteristics of the pseudo-feature cluster, such as rules based on shape template matching, rules based on texture analysis or rules based on machine learning classification, which can be implemented by rule switching or model reconstruction, with the aim of switching to a rule type that is more suitable for the current situation when the impact pattern changes significantly.

[0133] In some preferred embodiments, the system first obtains current external environmental condition information, such as light intensity, temperature, and humidity, as well as information on the characteristics of objects near the finger acquired through sensors or image analysis, such as sweat and a small amount of dust. Based on this information, the system determines the coexistence of sweat and dust. Since the result of the attachment status determination is yes, the system determines corresponding basic impact patterns for both sweat and dust based on their characteristic information. For example, the basic impact pattern for sweat may be a fuzzy function that simulates light scattering and absorption, while the basic impact pattern for dust may be a noise model that simulates light spots or shadows. The system then obtains a preset superposition logic, such as a weighted average model, and combines the basic impact patterns for sweat and dust according to this superposition logic to form composite impact pattern information. This composite impact pattern information describes the combined impact of sweat and dust on optical imaging. Finally, the system adjusts the preset pseudo-feature clustering morphology rules based on the current external environmental condition information (e.g., high humidity may enhance the impact of sweat) and this composite impact pattern information. For example, if the preset rule is a template based on shape matching, the system can fuzzify the template and superimpose a noise pattern based on the composite impact pattern information. Alternatively, if the composite impact pattern information indicates that the impact is very complex, the system can adjust the rule structure and switch to a rule based on texture analysis or machine learning classification. The current comparison rule generated in this way is used to compare with the extracted tissue morphological features.

[0134] Optional, combined Figure 6 As shown, S5533 combines all basic impact patterns according to the preset superposition logic to form composite impact pattern information, including:

[0135] A1, obtain the preset superposition logic;

[0136] A2, obtaining coupling indication information between all basic impact modes;

[0137] A3, selecting a coupling effect adjustment rule corresponding to the coupling indication information from a preset coupling effect adjustment rule set;

[0138] A4, apply the coupling effect adjustment rules to adjust the superposition logic;

[0139] A5, using the adjusted superposition logic, combines all basic impact patterns to form composite impact pattern information.

[0140] Among them, the preset superposition logic refers to an initial rule or model determined before the combination of basic influence patterns, which is used to merge or integrate the influences of multiple basic influence patterns. It can be implemented by linear superposition, weighted average, logic gate combination, etc., and its purpose is to provide a preliminary combination framework; basic influence pattern refers to a specific pattern caused by a single type of attachment or a specific state parameter of a single attachment, which affects optical fingerprint imaging. It can be represented by a mathematical model, a lookup table or an empirical curve, and its purpose is to quantify the independent influence of different attachments on imaging; coupling indication information refers to information used to characterize the degree or nature of the interaction between all basic influence patterns, which can be represented by one or a group of numerical values, a flag bit or a descriptor, and its purpose is to indicate whether the combined effect of the basic influence patterns deviates from simple superposition, and the characteristics of the deviation; the preset coupling effect adjustment rule set refers to the preset A rule base or model set is first established, which contains specific rules or parameters designed for different types of coupling indication information and used to adjust the overlay logic. It can be stored in a lookup table, decision tree or parameterized model set, with the purpose of providing a basis for adjusting the overlay logic; the coupling effect adjustment rule refers to a specific rule or parameter selected from a preset coupling effect adjustment rule set and matching the current coupling indication information. It can be an adjustment coefficient, a correction function or a logical judgment condition, with the purpose of correcting the overlay logic according to the specific coupling situation; the composite influence pattern information refers to the comprehensive information formed by combining all basic influence patterns according to the adjusted overlay logic, which is used to more accurately simulate the influence of interference from neighboring fingers on the fingerprint image. It can be represented by a comprehensive influence model, a corrected image template or a set of correction parameters, with the purpose of providing a more accurate interference model.

[0141] In some preferred embodiments, for example, assume there are two fundamental impact patterns: an optical scattering pattern caused by a small amount of sweat on the finger surface, and an edge shadow pattern caused by a slight lift of the finger edge. The preset overlay logic can simply perform pixel-by-pixel summation of the impact images of these two patterns. In practice, sweat may alter the contact state between the finger and the tablet, thereby affecting the shadow characteristics caused by the lift, and vice versa, resulting in a coupling effect. Through experiments or simulations, the actual effect of these two patterns acting together can be obtained and compared with the expected effect from simple addition to obtain coupling indication information, for example, indicating that sweat enhances the contrast of the edge shadow. The preset set of coupling effect adjustment rules can include a rule: If sweat enhances the edge shadow contrast, a coefficient related to the sweat pattern strength is added to the overlay logic to amplify the impact of the edge shadow pattern. After applying this rule to adjust the overlay logic, the new overlay logic can become: the composite impact equals the sweat pattern impact plus the edge shadow pattern impact multiplied by an adjustment factor based on the sweat enhancement coefficient. Using this adjusted overlay logic, the sweat pattern and the edge shadow pattern are combined to form more accurate composite impact pattern information.

[0142] Optional, combined Figure 7 As shown, the step of A2 obtaining coupling indication information between all basic impact modes includes:

[0143] A21, obtain the expected combined effect characteristics of all basic impact modes determined under the preset benchmark coupling conditions;

[0144] A22, by analyzing the interactive characteristics of the basic influence patterns, obtain the actual combined effect characteristics when the basic influence patterns act together;

[0145] A23, compare the actual combination effect characteristics with the expected combination effect characteristics to determine the effect deviation;

[0146] A24, determining whether the effect deviation exceeds a preset deviation threshold, and whether the pattern of the effect deviation deviates from any pattern in a preset set of known coupling patterns;

[0147] A25, if the judgment result is yes, generating coupling indication information based on the deviation condition and the deviation condition.

[0148] Among them, the basic influence pattern refers to the specific optical or physical effect pattern that affects the target finger fingerprint image caused by a single type of adjacent finger attachment or the characteristics of a single attachment under a specific state. It can be described in the form of a mathematical model, a parameter set, or a lookup table. Coupling indication information refers to information that quantitatively describes the degree and nature of the interaction between all basic influence patterns and the nonlinear superposition effect produced. It can be represented in the form of numerical values, vectors, identifiers, or structured data. The coupling effect adjustment rule set refers to a pre-established rule set used to guide how to modify or optimize the superposition method of basic influence patterns based on the coupling indication information. It can include a series of conditional judgment statements, parameter adjustment formulas, algorithm selection logic, or lookup tables. Superposition logic refers to the method or algorithm used to combine information from multiple basic influence patterns to predict or simulate the composite effect produced by their joint action. It can take the form of linear superposition, nonlinear function combination, model-based fusion algorithm, or machine learning model. Composite influence pattern information refers to pattern information that comprehensively reflects the overall impact of all basic influence patterns acting together on the target finger fingerprint image after considering the mutual coupling effect. It can be represented in the form of superimposed influence parameters, simulated images, correction coefficients, or prediction models.

[0149] In one embodiment, obtaining coupling indication information between all basic influence patterns may include: determining, through experiments or simulations, the expected combined effect characteristics of basic influence patterns M1 and M2 under preset baseline coupling conditions, such as the expected image artifact intensity after superposition, I_expected. By analyzing the interaction characteristics of M1 and M2, such as the effect of sweat on the optical properties of particles, the actual combined effect characteristics of their combined effect are obtained, such as the actual image artifact intensity after superposition, I_actual. I_actual is compared with I_expected to determine an effect deviation Delta_I = I_actual - I_expected. Determining whether Delta_I exceeds a preset deviation threshold, such as 10%, and whether the pattern of Delta_I (e.g., the spatial distribution or spectral characteristics of the artifact) deviates from any pattern in a preset set of known coupling patterns. If the deviation exceeds the threshold and the pattern deviates from a known pattern, generating coupling indication information based on the magnitude of the deviation and the pattern of the deviation, such as an identifier or parameter indicating "sweat enhances particle reflection." A coupling effect adjustment rule corresponding to the coupling indication is selected from a preset set of coupling effect adjustment rules. For example, when "sweat enhances granular reflexes," the influence weight of M2 is increased by a coefficient K, or a nonlinear function f(M1, M2) is used instead of the direct M1+M2 combination. This rule is applied to adjust the superposition logic, for example, adjusting the original superposition logic L(M1, M2) = M1 + M2 to L'(M1, M2) = M1 + K* M2 or L''(M1, M2) = f(M1, M2). Finally, the adjusted superposition logic L' or L'' is used to combine the basic influence patterns M1 and M2 to form the adjusted composite influence pattern information.

[0150] Optional, combined Figure 8 As shown, the step A24 of determining whether the mode of the effect deviation deviates from any mode in the preset known coupling mode set includes:

[0151] A241, performs component analysis on the effect deviation pattern, taking the component that is consistent with the known coupling pattern set as the first effect component, and the component that is different from the known coupling pattern set as the second effect component;

[0152] A242, obtaining preset component impact quantification rules;

[0153] A243, applying the component influence quantification rule to determine the influence degree corresponding to the first effect component and the influence degree corresponding to the second effect component;

[0154] A244, according to the preset attribution judgment logic, based on the influence of the first effect component and the influence of the second effect component, determines whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set.

[0155] Among them, component analysis of the pattern of effect deviation refers to decomposing the overall pattern of effect deviation into different parts that constitute it, which can be achieved by using signal decomposition technology, pattern decomposition algorithm or eigenvector decomposition and other methods, and its purpose is to identify the known pattern components and unknown pattern components contained in the effect deviation; wherein, the first effect component refers to the part of the effect deviation pattern that has similarity or matching with any pattern in the preset known coupling pattern set, and its purpose is to identify the known components in the effect deviation; wherein, the second effect component refers to the part of the effect deviation pattern that is different from any pattern in the preset known coupling pattern set, and its purpose is to identify the unknown components in the effect deviation; wherein, the component influence quantification rule refers to the rule used to measure the first effect component and the second effect component. The rule for the proportion or contribution of the component in the overall effect deviation can be implemented by measurement methods such as energy, amplitude, correlation or eigenvector distance, and its purpose is to quantify the influence of different components on the overall characteristics of the effect deviation; among them, the influence degree refers to the numerical value obtained by applying the component influence quantification rule, which reflects the importance or contribution of the component in the effect deviation, and its purpose is to provide a quantitative basis; among them, the attribution judgment logic refers to the rule or algorithm used to comprehensively consider the influence of the first effect component and the influence of the second effect component, and judge whether the effect deviation pattern belongs to the known coupling pattern set based on this. It can be implemented by methods based on threshold comparison, proportion analysis or classifier model, and its purpose is to make a final judgment based on the quantitative results.

[0156] In some preferred embodiments, specifically, it is assumed that the pattern of the effect deviation can be represented as a signal vector. The pattern of the effect deviation is subjected to component analysis, and a sparse representation method can be used to project the signal vector onto the basis vector corresponding to the known coupling mode set. The projected part is used as the first effect component, and the residual part after the projection is used as the second effect component. A preset component influence quantization rule is obtained, which can be defined as the energy of the component vector (i.e., the sum of the squares of the vector elements). The component influence quantization rule is applied to calculate the energy of the first effect component vector as the influence degree corresponding to the first effect component, and calculate the energy of the second effect component vector as the influence degree corresponding to the second effect component. According to the preset attribution judgment logic, the logic can set a threshold. For example, when the ratio of the influence degree of the second effect component to the influence degree of the first effect component exceeds a preset threshold, it is determined that the pattern of the effect deviation does not conform to any pattern in the known coupling mode set.

[0157] Optional, combined Figure 9 As shown, the step of determining whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set based on the influence degree of the first effect component and the influence degree of the second effect component according to the preset attribution determination logic includes:

[0158] A2441, Acquiring Physical Characteristics of Proximal Fingerprints and Imaging Parameters of Optical Fingerprint Recognition Systems;

[0159] A2442, determining judgment parameters based on physical characteristics and imaging parameters;

[0160] A2443, based on the judgment parameters and the influence of the first effect component and the influence of the second effect component, determines whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set.

[0161] Among them, the physical characteristics of the neighboring fingers refer to the inherent properties of the fingers, such as size, shape, skin texture, and flexibility, which can be obtained by image analysis, depth sensor measurement or user input; the imaging parameters of the optical fingerprint recognition system refer to the configuration or properties of the light source, such as the wavelength, illumination angle, sensor type, resolution, gain setting, etc. that affect the image acquisition process, which can be obtained by reading system configuration information or sensor metadata; the judgment parameters refer to indicators calculated or determined based on physical characteristics and imaging parameters, which are used to quantify the interference effect of neighboring fingers, such as the degree of occlusion of the light path by the neighboring fingers, the intensity distribution of reflected or scattered light, the local deformation transfer coefficient caused by finger pressure, etc., which can be determined by methods such as establishing physical models, optical simulation or data-driven regression analysis; the attribution judgment logic refers to a preset rule or algorithm used to comprehensively consider multiple input factors and output a judgment result on whether the effect deviation pattern belongs to a known coupling mode set. This logic can be implemented in various forms such as threshold comparison, machine learning models, and expert system rules.

[0162] In some preferred embodiments, in an optical fingerprint recognition system with simultaneous multi-finger presses, while capturing an image of the target finger, the system can simultaneously acquire image information of neighboring fingers. By analyzing the images of the neighboring fingers, the system can estimate the relative position of the neighboring fingers to the target finger, the contact area, and the morphological characteristics of the finger edges, which can be used as part of the physical characteristics of the neighboring fingers. Simultaneously, the system reads the current operating status of the optical fingerprint recognition system and obtains imaging parameters such as light source type, illumination angle, sensor model, and gain setting. Based on these physical characteristics and imaging parameters, the system calculates judgment parameters. For example, if the neighboring finger is very close to the target finger and the contact area is large, the system can determine a higher potential light obstruction and deformation transfer risk value as a judgment parameter. If the light source is of a specific wavelength and the finger skin is dry, the system can determine a higher likelihood of reflected light interference as a judgment parameter. Based on the influence degree of the first effect component (reflecting the degree of similarity to the known pattern) and the influence degree of the second effect component (reflecting the degree of difference from the known pattern) of the effect deviation obtained through component analysis and influence quantification, the system makes a final judgment based on the preset attribution judgment logic. For example, the logic can be a decision tree or a support vector machine model, whose inputs include the influence of the first effect component, the influence of the second effect component, and the calculated judgment parameter. If the influence of the second effect component is high and the judgment parameter indicates a high risk of interference from a neighboring finger, the judgment logic outputs "does not conform to any of the known coupling patterns." Conversely, if the influence of the second effect component is high but the judgment parameter indicates a low risk of interference from a neighboring finger, the judgment logic may output "conform to a pattern in the known coupling pattern set," indicating that the effect deviation may be caused by other known reasons.

[0163] An optical fingerprint recognition system for performing optical fingerprint recognition, combined with Figure 10 As shown, the optical fingerprint recognition system 1 includes:

[0164] A fingerprint image acquisition module 11 is used to acquire a fingerprint image of a target finger from an optical fingerprint recognition system;

[0165] A detail information extraction module 12 is used to extract preliminary fingerprint details based on the fingerprint image and generate preliminary fingerprint detail information;

[0166] The suspicious fingerprint screening module 13 is used to screen out the fingerprint details that are located in the edge area of ​​the fingerprint image and have suspicious morphology in the preliminary fingerprint detail information, and generate the suspicious fingerprint detail information;

[0167] The neighboring finger image acquisition module 14 is used to acquire the neighboring finger image information of the fingers adjacent to the target finger;

[0168] Interference correlation analysis module 15, used to analyze the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information, and generate corresponding interference cause correlation analysis results;

[0169] The recognition process correction module 16 is used to correct the suspicious fingerprint details during the fingerprint recognition process based on the interference cause correlation analysis result, so as to correct the fingerprint feature extraction error caused by the interference of adjacent fingers.

[0170] The fingerprint image acquisition module is a hardware or software unit used to capture raw image data of the target finger from the optical fingerprint recognition system, providing raw input for subsequent processing. The detail information extraction module is a hardware or software unit used to perform preliminary analysis of the acquired fingerprint image, identifying and extracting fingerprint detail features such as ridges, endpoints, and bifurcation points, with the goal of generating fingerprint detail data for further processing and analysis. The suspicious fingerprint screening module is a hardware or software unit used to identify and mark detail features located at the edge of the fingerprint image and exhibiting abnormal or irregular morphology within the initially extracted fingerprint detail information, with the goal of narrowing the potential interference area and improving analysis efficiency. The neighboring finger image acquisition module is a hardware or software unit used to acquire image data or related information of neighboring fingers simultaneously pressed on the photosensitive plate with the target finger, with the goal of providing reference information for inter-finger interference analysis. The interference correlation analysis module is a hardware or software unit used to evaluate whether there is a causal or correlation relationship between the suspicious fingerprint detail and the image information of neighboring fingers, with the goal of determining whether the suspicious detail is caused by interference from neighboring fingers. The recognition process correction module refers to a hardware or software unit used to adjust or correct the processing of suspicious fingerprint details in the fingerprint recognition process based on the results of interference correlation analysis. Its purpose is to eliminate or reduce the impact of interference from neighboring fingers on the final recognition results.

[0171] In some preferred embodiments, the optical fingerprint recognition system can be implemented as follows. The fingerprint image acquisition module can include a large-area optical photosensitive plate and associated image acquisition circuitry for simultaneously capturing image data from multiple fingers. The minutiae extraction module can be an image processor that runs standard fingerprint image preprocessing and feature extraction algorithms to generate a preliminary fingerprint minutiae list containing information such as the location, type, and orientation of each minutiae. The suspicious fingerprint screening module can utilize a post-processing program executed by the image processor. This program first determines the effective area outline of the fingerprint image and defines an edge region near the outline. Then, for each preliminary fingerprint minutiae within the edge region, its local image quality score and morphological regularity score are calculated. Minutiae with scores below a preset threshold are marked as suspicious, generating a list of suspicious fingerprint minutiae information. The neighboring finger image acquisition module can utilize image data from other fingers simultaneously acquired by the fingerprint image acquisition module, or utilize additional sensors to obtain the position and general morphological information of neighboring fingers. The interference correlation analysis module can be an independent analysis unit or an algorithm executed by the main processor that compares the position and morphological information of suspicious fingerprint minutiae with that of neighboring fingers. For example, if a suspicious detail is located at the edge of the target finger image and close to the adjacent finger, and its morphological features are consistent with the preset artifact pattern caused by the interference of the adjacent finger, then the suspicious detail is judged to be associated with the adjacent finger interference. The recognition process correction module can be a decision-making unit that adjusts the subsequent fingerprint feature comparison or recognition algorithm based on the output of the interference association analysis module. For example, for pseudo details determined to be caused by the interference of the adjacent finger, the recognition process correction module can instruct the comparison algorithm to reduce the weight of these details when calculating the similarity, or directly remove them from the feature template. For real details that are determined to have position or direction deviations due to interference, the recognition process correction module can attempt to correct their parameters.

[0172] Through the above technical solution, the present application can specifically identify and analyze the impact of optical or physical interference caused by neighboring fingers on the fingerprint image of the target finger. By correlating suspicious details in the edge area of ​​the fingerprint image with the image information of the neighboring fingers, it can be determined whether these suspicious details are caused by interference. Based on the results of this correlation analysis, the system can correct the feature extraction and comparison in the fingerprint recognition process, such as ignoring pseudo feature points or correcting damaged real features. This effectively avoids misjudging interference signals as fingerprint features, or affecting recognition due to distortion of real features due to interference, thereby improving the recognition accuracy of the optical fingerprint recognition system in scenarios where multiple fingers are input simultaneously, and reducing the false acceptance rate and false rejection rate.

[0173] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. An optical fingerprint recognition method, characterized in that: include: acquiring a fingerprint image of a target finger from an optical fingerprint recognition system; Performing preliminary fingerprint detail extraction based on the fingerprint image to generate preliminary fingerprint detail information; Screening out fingerprint details with suspicious morphology located in the edge area of ​​the fingerprint image from the preliminary fingerprint detail information to generate suspicious fingerprint detail information; Acquire image information of fingers adjacent to the target finger; Analyzing the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information to generate a corresponding interference cause correlation analysis result; the step of analyzing the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information to generate a corresponding interference cause correlation analysis result includes: identifying a fingerprint detail cluster area corresponding to the suspicious fingerprint detail information in an edge area of ​​the fingerprint image; extracting tissue morphological features of the fingerprint detail cluster area; obtaining position information of the adjacent finger based on the adjacent finger image information; judging whether the fingerprint detail cluster area is within the potential optical influence range of the adjacent finger based on the position information, and generating a position judgment result; comparing the tissue morphological features with preset pseudo-feature cluster morphological rules to generate a morphological comparison result; the preset pseudo-feature cluster morphological rules refer to a pre-established set of typical morphological manifestation patterns of pseudo-features that may be caused by interference from adjacent fingers; performing a correlation analysis based on the morphological comparison result and the position judgment result, and determining that a correlation exists when both the morphological comparison result and the position judgment result indicate correlation with interference from the adjacent finger, and generating the interference cause correlation analysis result based on the correlation analysis result; Based on the interference cause correlation analysis result, the suspicious fingerprint detail information in the fingerprint recognition process is corrected to correct the fingerprint feature extraction error caused by the interference of adjacent fingers.

2. The optical fingerprint recognition method according to claim 1, characterized in that: The step of screening out fingerprint details with suspicious morphology located in the edge area of ​​the fingerprint image from the preliminary fingerprint detail information and generating suspicious fingerprint detail information comprises: Obtaining a fingerprint area contour line of the fingerprint image, and inwardly expanding a preset pixel width based on the fingerprint area contour line to determine an edge area of ​​the fingerprint image; In the edge area, performing local image quality scoring and geometric morphology analysis on each fingerprint detail to generate a corresponding suspicion score; The fingerprint details corresponding to the suspicion scores exceeding a preset threshold are organized into the suspicious fingerprint detail information.

3. The optical fingerprint recognition method according to claim 2, characterized in that: The step of comparing the tissue morphological features with preset pseudo-feature cluster morphological rules to generate a morphological comparison result includes: Acquiring external environmental condition information corresponding to when the optical fingerprint recognition system collects the fingerprint image and characteristic information of an attachment corresponding to the adjacent finger; determining whether a preset pseudo-feature cluster morphology rule is adapted to the pseudo-feature cluster morphology caused by the concealed optical effect of the adjacent finger under the conditions corresponding to the external environmental condition information and the attachment characteristic information, and generating a rule adaptability determination result; If the rule adaptability judgment result is no, adjusting and updating the pseudo-feature cluster morphology rule according to the external environment condition information and the attachment characteristic information to generate a current comparison rule; If the rule adaptability judgment result is yes, then directly use the pseudo-feature cluster morphology rule as the current comparison rule; The tissue morphological features are compared with the current comparison rules to generate the morphological comparison results.

4. The optical fingerprint recognition method according to claim 3, characterized in that: The step of adjusting and updating the pseudo-feature cluster morphology rule according to the external environmental condition information and the attachment characteristic information to generate the current comparison rule comprises: Based on the attachment characteristic information, determining whether at least two types of attachments coexist on the surface of the adjacent finger, or whether characteristic parameters of a single type of attachment on the surface of the adjacent finger change over time, and generating an attachment status determination result; If the result of the attachment state determination is yes, determining a corresponding basic influence mode according to the attachment characteristic information for each attachment type in the coexistence of the at least two types of attachments, or for current characteristics of a single type of attachment in the single type of attachment situation; Combining all the basic impact patterns according to a preset superposition logic to form composite impact pattern information; According to the external environmental condition information and the composite impact pattern information, the parameters of the preset pseudo-feature cluster morphology rule are adjusted, or the structure of the preset pseudo-feature cluster morphology rule is adjusted to generate a current comparison rule.

5. The optical fingerprint recognition method according to claim 4, characterized in that: The step of combining all the basic impact patterns according to the preset superposition logic to form composite impact pattern information includes: Get the preset overlay logic; Acquiring coupling indication information between all the basic impact modes; Selecting a coupling effect adjustment rule corresponding to the coupling indication information from a preset coupling effect adjustment rule set; Applying the coupling effect adjustment rule to adjust the superposition logic; All the basic impact patterns are combined using the adjusted superposition logic to form the composite impact pattern information.

6. The optical fingerprint recognition method according to claim 5, characterized in that: The step of obtaining coupling indication information between all the basic influence modes includes: Obtaining expected combined effect characteristics of all the basic impact modes determined under preset benchmark coupling conditions; By analyzing the interactive characteristics of the basic influence modes, the actual combined effect characteristics of the basic influence modes when acting together are obtained; comparing the actual combination effect characteristics with the expected combination effect characteristics to determine effect deviation; Determining whether the effect deviation exceeds a preset deviation threshold, and whether a pattern of the effect deviation deviates from any pattern in a preset set of known coupling patterns; If the judgment result is yes, the coupling indication information is generated based on the deviation condition and the deviation condition.

7. The optical fingerprint recognition method according to claim 6, characterized in that: The step of determining whether the mode of the effect deviation deviates from any mode in a preset set of known coupling modes includes: Performing component analysis on the pattern of the effect deviation, taking the component consistent with the known coupling pattern set as the first effect component, and taking the component different from the known coupling pattern set as the second effect component; Obtain preset component impact quantification rules; Applying the component influence quantification rule to determine the influence degree corresponding to the first effect component and the influence degree corresponding to the second effect component; According to a preset attribution determination logic, based on the influence degree of the first effect component and the influence degree of the second effect component, it is determined whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set.

8. The optical fingerprint recognition method according to claim 7, characterized in that: The step of determining whether the pattern of the effect deviation does not conform to any pattern in the set of known coupling patterns based on the influence degree of the first effect component and the influence degree of the second effect component according to a preset attribution determination logic includes: Acquiring physical properties of the adjacent finger and imaging parameters of the optical fingerprint recognition system; determining a determination parameter based on the physical characteristic and the imaging parameter; According to the determination parameter, based on the influence degree of the first effect component and the influence degree of the second effect component, it is determined whether the pattern of the effect deviation does not conform to any pattern in the known coupling pattern set.

9. An optical fingerprint recognition system for performing optical fingerprint recognition, characterized in that: include: A fingerprint image acquisition module, used to acquire a fingerprint image of a target finger from an optical fingerprint recognition system; A detail information extraction module, configured to extract preliminary fingerprint detail information based on the fingerprint image and generate preliminary fingerprint detail information; a suspicious fingerprint screening module, configured to screen out fingerprint details of suspicious morphology located in the edge area of ​​the fingerprint image from the preliminary fingerprint detail information, and generate suspicious fingerprint detail information; A neighboring finger image acquisition module, configured to acquire image information of neighboring fingers of the target finger; The interference correlation analysis module is used to analyze the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information, and generate a corresponding interference cause correlation analysis result; the steps of analyzing the interference cause correlation between the suspicious fingerprint detail information and the adjacent finger image information and generating a corresponding interference cause correlation analysis result include: identifying the fingerprint detail cluster area corresponding to the suspicious fingerprint detail information in the edge area of ​​the fingerprint image; extracting the tissue morphological characteristics of the fingerprint detail cluster area; obtaining the position information of the adjacent finger based on the adjacent finger image information; judging the fingerprint according to the position information. Whether the detail cluster area is within the potential optical influence range of the adjacent finger is determined, and a position judgment result is generated; the tissue morphological features are compared with preset pseudo-feature cluster morphological rules to generate a morphological comparison result; the preset pseudo-feature cluster morphological rules refer to a pre-established set of typical morphological manifestation patterns of pseudo-features that may be caused by interference from adjacent fingers; a correlation analysis is performed based on the morphological comparison result and the position judgment result. When the morphological comparison result and the position judgment result both indicate correlation with interference from the adjacent finger, a correlation is determined to exist, and the interference cause correlation analysis result is generated based on the correlation analysis result; The recognition process correction module is used to correct the suspicious fingerprint detail information during the fingerprint recognition process based on the interference cause correlation analysis result, so as to correct the fingerprint feature extraction error caused by the interference of adjacent fingers.

Images