Fingerprint comparison method and system based on multi-feature fusion

Abstract

The invention provides a fingerprint comparison method and system based on multi-feature fusion. The method comprises the following steps: obtaining a fingerprint image to be compared, preprocessing the fingerprint image to be compared to obtain a standard fingerprint image, extracting image quality evaluation data, processing the image quality evaluation data to obtain an image quality evaluation index, carrying out threshold comparison, and if the standard fingerprint image is judged to be qualified, carrying out feature extraction; according to the method, global feature data and local feature data are obtained, the global feature data and the local feature data are compared with fingerprint feature data in a preset fingerprint database to obtain corresponding similarity scores, weighted averaging processing is carried out to obtain comparison similarity evaluation indexes, and finally a comparison result is obtained through threshold comparison; according to the method and the device, the accuracy and the reliability of fingerprint comparison based on multi-feature fusion are improved through standard fingerprint image evaluation, multi-feature extraction and dynamic weight-based fusion processing.

Description

Technical Field

[0001] This application relates to the field of biometric technology, and more particularly, to a fingerprint comparison method and system based on multi-feature fusion. Background Art

[0002] Fingerprint recognition, as a mature and widely used biometric technology, plays an important role in many fields such as security, finance, and unlocking of electronic devices. Traditional fingerprint comparison methods often rely only on a single feature, such as the minutiae feature of a fingerprint. However, due to the fact that fingerprints may be affected by various factors during the acquisition process, such as the pressing force of the finger, the placement angle, and the accuracy of the acquisition device, it is easy to result in uneven fingerprint image quality. Relying solely on a single feature for comparison is prone to misjudgment, and the recognition accuracy and reliability are difficult to meet the security and application requirements.

[0003] In view of the above problems, there is an urgent need for effective technical solutions. Summary of the Invention

[0004] The purpose of this application is to provide a fingerprint comparison method and system based on multi-feature fusion, which can evaluate standard fingerprint images, extract multi-features, and perform fusion processing based on dynamic weights, thereby improving the accuracy and reliability of fingerprint comparison based on multi-feature fusion.

[0005] This application also provides a fingerprint comparison method based on multi-feature fusion, including the following steps: Obtain the fingerprint image to be compared and perform preprocessing to obtain a standard fingerprint image; Extract according to the standard fingerprint image to obtain image quality evaluation data, perform processing to obtain an image quality evaluation index, and compare it with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint image; If the qualified status of the standard fingerprint image is qualified, perform feature extraction according to the standard fingerprint image to obtain global feature data and local feature data; Compare the global feature data and the local feature data with the fingerprint feature data in the preset fingerprint database respectively to obtain corresponding similarity scores, and perform weighted average processing to obtain a comparison similarity evaluation index; Compare the comparison similarity evaluation index with a preset comparison similarity evaluation threshold to obtain a comparison result.

[0006] Optionally, in the fingerprint comparison method based on multi-feature fusion described in this application, the step of extracting according to the standard fingerprint image to obtain image quality evaluation data, performing processing to obtain an image quality evaluation index, and comparing it with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint image includes: Extract the standard fingerprint image to obtain image quality evaluation data, including image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data; Perform weighted summation processing on the image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data to obtain an image quality evaluation index; Compare the image quality evaluation index with a preset image quality evaluation threshold; If it is less than or equal to the preset image quality evaluation threshold, determine that the standard fingerprint image is unqualified; If it is greater than the preset image quality evaluation threshold, determine that the standard fingerprint image is qualified.

[0007] Optionally, in the fingerprint comparison method based on multi-feature fusion described in this application, if the qualified state of the standard fingerprint image is qualified, perform feature extraction on the standard fingerprint image to obtain global feature data and local feature data, including: If the qualified state of the standard fingerprint image is qualified, perform feature extraction on the standard fingerprint image to obtain global feature data and local feature data; The global feature data includes direction field feature data and frequency feature data; The local feature data includes minutiae feature data and local texture feature data.

[0008] Optionally, in the fingerprint comparison method based on multi-feature fusion described in this application, it further includes: Obtain the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period; Perform weighted summation processing on the accuracy rate, false recognition rate, rejection rate, and equal error rate in combination with corresponding preset weight values to obtain an accuracy evaluation index for fingerprint comparison; Process the accuracy evaluation index and a preset accuracy requirement index to obtain a comparison accuracy deviation rate; Compare the comparison accuracy deviation rate with a preset comparison accuracy level threshold, and obtain an accuracy evaluation level according to the threshold range into which the comparison accuracy deviation rate falls; Query a preset feature weight value list according to the accuracy evaluation level to obtain dynamic feature weight values, including a first feature weight value for global feature data and a second feature weight value for local feature data.

[0009] Optionally, in the fingerprint comparison method based on multi-feature fusion described in this application, when comparing the global feature data and the local feature data with the fingerprint feature data in a preset fingerprint database respectively to obtain corresponding similarity scores and performing weighted average processing to obtain a comparison similarity evaluation index, including: Compare the global feature data and the local feature data with the fingerprint feature data in a preset fingerprint database respectively to obtain corresponding similarity scores, including a global feature similarity score and a local feature similarity score; Perform weighted average processing on the global feature similarity score and the local feature similarity score by combining the first feature weight value and the second feature weight value to obtain a comparison similarity evaluation index.

[0010] Optionally, in the fingerprint comparison method based on multi-feature fusion described in this application, the comparing the comparison similarity evaluation index with a preset comparison similarity evaluation threshold to obtain a comparison result includes: Compare the comparison similarity evaluation index with a preset comparison similarity requirement index to obtain a comparison similarity relative value; Compare the comparison similarity relative value with a preset comparison similarity evaluation threshold, where the preset comparison similarity evaluation threshold includes a first preset comparison similarity evaluation threshold and a second preset comparison similarity evaluation threshold; If it is less than or equal to the first preset comparison similarity evaluation threshold, it is determined that the comparison result is a comparison failure; If it is greater than the first preset comparison similarity evaluation threshold and less than or equal to the second preset comparison similarity evaluation threshold, it is determined that the comparison result is pending confirmation; If it is greater than the second preset comparison similarity evaluation threshold, it is determined that the comparison result is a comparison success.

[0011] Optionally, in the fingerprint comparison method based on multi-feature fusion described in this application, it further includes: Obtain the acquisition distortion evaluation data of the fingerprint image to be compared, including the pressing force, the placement angle, and the acquisition time; Process the pressing force with a preset minimum pressing force and a preset maximum pressing force to obtain a pressing force deviation rate; Compare the placement angle with preset placement angle parameter data to obtain a placement angle deviation rate; Compare the acquisition time with preset acquisition time parameter data to obtain an acquisition time deviation rate; Process the pressing force deviation rate, the placement angle deviation rate, and the acquisition time deviation rate by combining a preset pressure weight value, a preset placement angle weight value, and a preset acquisition time weight value to obtain a predicted image distortion index; Compare the predicted image distortion index with a preset image acquisition quality evaluation threshold; If the predicted image distortion index is less than or equal to the preset image acquisition quality evaluation threshold, it is determined that the image acquisition is effective; If the predicted image distortion index is greater than the preset image acquisition quality evaluation threshold, it is determined that the image acquisition is invalid.

[0012] In a second aspect, the present application provides a fingerprint comparison system based on multi-feature fusion. The system includes: a memory and a processor. The memory includes a program of a fingerprint comparison method based on multi-feature fusion. When the program of the fingerprint comparison method based on multi-feature fusion is executed by the processor, the following steps are implemented: Obtain the fingerprint images to be compared and perform preprocessing to obtain standard fingerprint images; Extract according to the standard fingerprint images to obtain image quality evaluation data, perform processing to obtain an image quality evaluation index, and perform threshold comparison with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint images; If the qualified status of the standard fingerprint images is qualified, perform feature extraction according to the standard fingerprint images to obtain global feature data and local feature data; Compare the global feature data and the local feature data with the fingerprint feature data in the preset fingerprint database respectively to obtain corresponding similarity scores, and perform weighted average processing to obtain a comparison similarity evaluation index; Perform threshold comparison between the comparison similarity evaluation index and a preset comparison similarity evaluation threshold to obtain a comparison result.

[0013] Optionally, in the fingerprint comparison system based on multi-feature fusion described in the present application, the extracting according to the standard fingerprint images to obtain image quality evaluation data, performing processing to obtain an image quality evaluation index, and performing threshold comparison with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint images includes: Extract the standard fingerprint images to obtain image quality evaluation data, including image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data; Perform weighted summation processing on the image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data to obtain an image quality evaluation index; Perform threshold comparison between the image quality evaluation index and a preset image quality evaluation threshold; If it is less than or equal to the preset image quality evaluation threshold, determine that the standard fingerprint images are unqualified; If it is greater than the preset image quality evaluation threshold, determine that the standard fingerprint images are qualified.

[0014] Optionally, in the fingerprint comparison system based on multi-feature fusion described in the present application, the if the qualified status of the standard fingerprint images is qualified, perform feature extraction according to the standard fingerprint images to obtain global feature data and local feature data includes: If the qualified status of the standard fingerprint image is qualified, the standard fingerprint image is subjected to feature extraction to obtain global feature data and local feature data; The global feature data includes direction field feature data and frequency feature data; The local feature data includes minutiae feature data and local texture feature data.

[0015] As can be seen from the above, a fingerprint comparison method and system based on multi-feature fusion provided by the present application realizes improving the accuracy and reliability of fingerprint comparison based on multi-feature fusion through evaluating the standard fingerprint image, multi-feature extraction, and fusion processing based on dynamic weights.

[0016] Other features and advantages of the present application will be described in the subsequent specification, and, in part, will become apparent from the specification or be understood by implementing the embodiments of the present application. The objectives and other advantages of the present application can be realized and obtained by the structures specifically pointed out in the written specification and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of a fingerprint comparison method based on multi-feature fusion provided by an embodiment of the present application; Figure 2 It is a flowchart of obtaining the qualified status of the standard fingerprint image of a fingerprint comparison method based on multi-feature fusion provided by an embodiment of the present application; Figure 3 It is a flowchart of obtaining the comparison similarity evaluation index of a fingerprint comparison method based on multi-feature fusion provided by an embodiment of the present application; Figure 4 It is a flowchart of obtaining the comparison result of a fingerprint comparison method based on multi-feature fusion provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and illustrated in the accompanying 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 accompanying drawings is not intended to limit the scope of the present application to be protected, but only represents the selected embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0020] 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 and explained in subsequent drawings. At the same time, in the description of the present application, terms such as "first" and "second" are only used for differential description and cannot be understood as indicating or implying relative importance.

[0021] Please refer to Figure 1 , Figure 1 which is a flowchart of a fingerprint comparison method based on multi-feature fusion in some embodiments of the present application. This fingerprint comparison method based on multi-feature fusion is used in terminal devices, such as computer, mobile phone terminals, etc. This fingerprint comparison method based on multi-feature fusion includes the following steps: S11. Obtain the fingerprint image to be compared, and perform preprocessing to obtain a standard fingerprint image; S12. Extract according to the standard fingerprint image to obtain image quality evaluation data, and perform processing to obtain an image quality evaluation index, and perform a threshold comparison with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint image; S13. If the qualified status of the standard fingerprint image is qualified, perform feature extraction according to the standard fingerprint image to obtain global feature data and local feature data; S14. Compare the global feature data and the local feature data with the fingerprint feature data in the preset fingerprint library respectively to obtain corresponding similarity scores, and perform weighted average processing to obtain a comparison similarity evaluation index; S15. Perform a threshold comparison between the comparison similarity evaluation index and a preset comparison similarity evaluation threshold to obtain a comparison result.

[0022] It should be noted that, in order to perform fingerprint comparison accurately and efficiently, the collected fingerprint images to be compared are first subjected to graying, filtering to remove noise, binarization, and thinning preprocessing to obtain standard fingerprint images. To ensure that the quality of the collected images meets the standards, image quality evaluation data is extracted and processed to obtain an image quality evaluation index, and then a threshold comparison is performed to obtain the qualified status of the standard fingerprint image. If it is unqualified, a warning reminder is output to prompt re-collection. If it is qualified, global feature extraction and local feature extraction are further performed to provide data support for multi-feature fusion. Then, the extracted global feature data and local feature data are respectively compared with the fingerprint feature data in the preset fingerprint database to obtain corresponding similarity scores. Through weighted average processing, a comparison similarity evaluation index is obtained. Finally, it is determined whether the fingerprint comparison is successful through threshold comparison.

[0023] Please refer to Figure 2 , Figure 2 is a flowchart of obtaining the qualified status of a standard fingerprint image for a fingerprint comparison method based on multi-feature fusion in some embodiments of the present application. According to an embodiment of the present invention, the extracting, obtaining image quality evaluation data, and processing according to the standard fingerprint image to obtain an image quality evaluation index, and performing a threshold comparison with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint image includes: S21. Extract the standard fingerprint image to obtain image quality evaluation data, including image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data; S22. Perform weighted summation processing on the image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data to obtain an image quality evaluation index; S23. Perform a threshold comparison between the image quality evaluation index and a preset image quality evaluation threshold; S24. If it is less than or equal to the preset image quality evaluation threshold, determine that the standard fingerprint image is unqualified; S25. If it is greater than the preset image quality evaluation threshold, determine that the standard fingerprint image is qualified.

[0024] It should be noted that before multi-feature extraction, the collected fingerprint images should be evaluated to avoid low-quality fingerprint images affecting feature extraction and fingerprint comparison results. By extracting image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data and performing weighted summation processing, an image quality evaluation index is obtained. Among them, the ridge integrity data is obtained by calculating the continuous length and breakage times of the ridges in the fingerprint image, and the corresponding weight values are obtained by querying the preset fingerprint comparison platform. Then, a threshold comparison is performed, and it is determined whether the standard fingerprint image is qualified according to the threshold comparison result.

[0025] According to an embodiment of the present invention, if the qualified state of the standard fingerprint image is qualified, feature extraction is performed on the standard fingerprint image to obtain global feature data and local feature data, including: If the qualified state of the standard fingerprint image is qualified, feature extraction is performed on the standard fingerprint image to obtain global feature data and local feature data; The global feature data includes orientation field feature data and frequency feature data; The local feature data includes minutiae feature data and local texture feature data.

[0026] It should be noted that in order to achieve multi-feature fusion processing, first, feature extraction is performed on the standard fingerprint image to obtain global feature data including orientation field feature data and frequency feature data. Among them, the orientation field feature data is obtained by calculating the gradients of the fingerprint image in the x and y directions using the Sobel operator, obtaining the gradient magnitude and direction of each pixel point in the image, and then dividing the fingerprint image into several appropriately sized blocks, usually blocks of 16×16 or 32×32 pixels. For each block, the gradient direction histogram of the pixel points therein is statistically analyzed to represent the dominant direction of the fingerprint ridges in the block, obtaining the orientation field of the fingerprint. The frequency feature data is obtained by filtering the fingerprint image using a Gabor filter. The frequency and direction parameters of the Gabor filter match the frequency and direction of the fingerprint ridges. Through filtering operations at different frequencies and directions, the responses of the fingerprint image at different frequencies are obtained, and then statistical analysis is performed on the filtered image, such as calculating the energy distribution on each frequency channel as the frequency feature of the fingerprint; local feature data including minutiae feature data and local texture feature data. Among them, the minutiae feature data is obtained by binarizing the fingerprint image, converting it into a black-and-white image to highlight the difference between the fingerprint ridges and the background, then thinning the binarized image to obtain a single-pixel-wide fingerprint ridge skeleton, and finally, on the thinned image, by analyzing the endpoint and bifurcation point structures of the ridges, the minutiae of the fingerprint are detected. The local texture feature data is obtained by dividing the fingerprint image into smaller local regions, such as small blocks of 8×8 or 16×16 pixels, and then for each small block, calculating its texture features, such as gray-level co-occurrence matrix (GLCM) features. The GLCM can describe the spatial relationship between different gray levels in the image, and contrast, correlation, energy, and entropy are extracted therefrom to characterize the characteristics of the local texture of the fingerprint.

[0027] According to an embodiment of the present invention, it further includes: Obtain the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period; Perform weighted summation processing on the accuracy rate, false recognition rate, rejection rate, and equal error rate in combination with the corresponding preset weight values to obtain an accuracy evaluation index for fingerprint comparison; Process the accuracy evaluation index and a preset accuracy requirement index to obtain a comparison accuracy deviation rate; Compare the comparison accuracy deviation rate with a preset comparison accuracy level threshold, and obtain an accuracy evaluation level according to the threshold range in which the comparison accuracy deviation rate falls; Query a preset list of characteristic weight values according to the accuracy evaluation level to obtain dynamic characteristic weight values, including a first characteristic weight value of global characteristic data and a second characteristic weight value of local characteristic data.

[0028] It should be noted that when performing feature fusion, if fixed characteristic weight values are used, the adaptability is insufficient. Therefore, in this embodiment, dynamic characteristic weight values are used. To obtain dynamic characteristic weight values, first, obtain the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period. Among them, the accuracy rate is the ratio of the number of correctly matched fingerprint pairs to the total number of comparison pairs. The false recognition rate is the probability of mis-matching fingerprints of different people as those of the same person. The rejection rate is the probability of mis-determining fingerprints of the same person as those of different people. The equal error rate is the error rate when the false recognition rate and the rejection rate are equal. Then, perform weighted summation processing in combination with the corresponding preset weight values to obtain an accuracy evaluation index for fingerprint comparison. Then, process the accuracy evaluation index and a preset accuracy requirement index to obtain a comparison accuracy deviation rate. The comparison accuracy deviation rate is the ratio of the absolute value of the difference between the accuracy evaluation index and the preset accuracy requirement index to the preset accuracy requirement index; then compare the comparison accuracy deviation rate with a preset comparison accuracy level threshold. In this embodiment, the preset comparison accuracy level threshold is set to (0, 0.35], (0.35, 0.55], (0.55, 1], corresponding to first-level accuracy, second-level accuracy, and third-level accuracy respectively, where the first-level accuracy level is the highest. For example, if the obtained accuracy evaluation index is 7 and the preset accuracy requirement index is 10, then 3 / 10 = 0.3 is the comparison accuracy deviation rate, and the accuracy evaluation level is determined to be first-level accuracy. Finally, query a preset list of characteristic weight values to obtain a first characteristic weight value of global characteristic data and a second characteristic weight value of local characteristic data, where the preset list of characteristic weight values is obtained by querying a preset fingerprint comparison platform; this method realizes: obtaining the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period, and performing weighted summation processing to obtain an accuracy evaluation index for fingerprint comparison; processing the accuracy evaluation index and a preset accuracy requirement index to obtain a comparison accuracy deviation rate, comparing it with a preset comparison accuracy level threshold, and obtaining an accuracy evaluation level according to the threshold range in which the comparison accuracy deviation rate falls; querying a preset list of characteristic weight values according to the accuracy evaluation level to obtain dynamic characteristic weight values.

[0029] Please refer to Figure 3 , Figure 3It is a flowchart for obtaining a comparison similarity evaluation index of a fingerprint comparison method based on multi-feature fusion in some embodiments of the present application. According to the embodiments of the present invention, the global feature data and the local feature data are respectively compared with the fingerprint feature data in a preset fingerprint database to obtain corresponding similarity scores, and weighted averaging is performed to obtain a comparison similarity evaluation index, including: S31. Respectively compare the global feature data and the local feature data with the fingerprint feature data in a preset fingerprint database to obtain corresponding similarity scores, including a global feature similarity score and a local feature similarity score; S32. Perform weighted averaging on the global feature similarity score and the local feature similarity score in combination with the first feature weight value and the second feature weight value to obtain a comparison similarity evaluation index.

[0030] It should be noted that when performing feature fusion, first, the global feature data and the local feature data are respectively compared with the fingerprint feature data in a preset fingerprint database to obtain corresponding similarity scores, and then weighted averaging is performed in combination with the first feature weight value and the second feature weight value to obtain a comparison similarity evaluation index; The calculation formula of the comparison similarity evaluation index is: ; Wherein, is the comparison similarity evaluation index, , are respectively the global feature similarity score and the local feature similarity score, , are respectively the first feature weight value and the second feature weight value.

[0031] Please refer to Figure 4 , Figure 4 is a flowchart for obtaining a comparison result of a fingerprint comparison method based on multi-feature fusion in some embodiments of the present application. According to the embodiments of the present invention, comparing the comparison similarity evaluation index with a preset comparison similarity evaluation threshold to obtain a comparison result includes: S41. Compare the comparison similarity evaluation index with a preset comparison similarity requirement index to obtain a relative comparison similarity value; S42. Compare the relative comparison similarity value with a preset comparison similarity evaluation threshold, wherein the preset comparison similarity evaluation threshold includes a first preset comparison similarity evaluation threshold and a second preset comparison similarity evaluation threshold; S43. If it is less than or equal to the first preset comparison similarity evaluation threshold, it is determined that the comparison result is a comparison failure; S44. If it is greater than the first preset comparison similarity evaluation threshold and less than or equal to the second preset comparison similarity evaluation threshold, then determine that the comparison result is pending confirmation; S45. If it is greater than the second preset comparison similarity evaluation threshold, then determine that the comparison result is a successful comparison.

[0032] It should be noted that in order to determine the fingerprint comparison result, first compare the obtained comparison similarity evaluation index with the preset comparison similarity requirement index to obtain the relative comparison similarity value. For example, if the comparison similarity evaluation index is 8 and the preset comparison similarity requirement index is 10, then 8 / 10 = 0.8 is the relative comparison similarity value. Then, compare it with the first preset comparison similarity evaluation threshold and the second preset comparison similarity evaluation threshold respectively to obtain the comparison result including comparison failure, pending confirmation of the comparison result, or successful comparison. In this embodiment, the preset comparison similarity evaluation threshold is set to (0, 0.7], (0.7, 0.8], (0.8, 1], corresponding to comparison failure, pending confirmation of the comparison result, and successful comparison respectively. For example, if the obtained relative comparison similarity value is 0.75, then determine that the comparison result is pending confirmation of the comparison result.

[0033] According to an embodiment of the present invention, it further includes: Obtain the acquisition distortion evaluation data of the fingerprint image to be compared, including the pressing force, placement angle, and acquisition time; Process the pressing force with the preset minimum pressing force and the preset maximum pressing force to obtain the pressing force deviation rate; Compare the placement angle with the preset placement angle parameter data to obtain the placement angle deviation rate; Compare the acquisition time with the preset acquisition time parameter data to obtain the acquisition time deviation rate; According to the pressing force deviation rate, placement angle deviation rate, and acquisition time deviation rate, combine the preset pressure weight value, preset placement angle weight value, and preset acquisition time weight value for processing to obtain the predicted image distortion index; Compare the predicted image distortion index with the preset image acquisition quality evaluation threshold; If the predicted image distortion index is less than or equal to the preset image acquisition quality evaluation threshold, then determine that the image acquisition is effective; If the predicted image distortion index is greater than the preset image acquisition quality evaluation threshold, then determine that the image acquisition is invalid.

[0034] It should be noted that during fingerprint acquisition, excessive pressing force can distort the fingerprint ridges, too little force can result in an incomplete fingerprint image, incorrect placement angle can cause the fingerprint image to be distorted, too short acquisition time cannot obtain a complete and clear fingerprint, and too long acquisition time can cause blurring due to finger movement; in order to evaluate the quality of fingerprint image acquisition, acquisition distortion evaluation data including pressing force, placement angle, and acquisition time are obtained, and processed in combination with corresponding preset parameter data to obtain a pressing force deviation rate, a placement angle deviation rate, and an acquisition time deviation rate. Among them, the pressing force deviation rate is the ratio of the difference between the pressing force and the minimum preset pressing force to the difference between the maximum preset pressing force and the minimum preset pressing force. If the obtained pressing force is less than or equal to the minimum pressing force, a warning is directly output. For example, if the pressing force is 7, the minimum preset pressing force is 5, and the maximum preset pressing force is 10, then (7 - 5) / (10 - 5) = 0.4 is the pressing force deviation rate. The placement angle deviation rate is the ratio of the absolute value of the difference between the placement angle and the preset placement angle parameter data to the preset placement angle parameter data, and the acquisition time deviation rate is the ratio of the absolute value of the difference between the acquisition time and the preset acquisition time parameter data to the preset acquisition time parameter data; according to the obtained pressing force deviation rate, placement angle deviation rate, and acquisition time deviation rate, combined with the preset pressure weight value, preset placement angle weight value, and preset acquisition time weight value, processing is carried out to obtain a predicted image distortion index; The calculation formula for the predicted image distortion index is: ; Wherein, is the predicted image distortion index, , , are the pressing force deviation rate, the placement angle deviation rate, and the acquisition time deviation rate respectively, , , are the preset pressure weight value, the preset placement angle weight value, and the preset acquisition time weight value respectively, is a preset feature coefficient (the feature coefficient is obtained by querying through a preset fingerprint comparison platform); finally, compare the obtained predicted image distortion index with a preset image acquisition quality evaluation threshold. In this embodiment, the preset image acquisition quality evaluation threshold is set to (0, 0.55] and (0.55, 1], corresponding to effective image acquisition and ineffective image acquisition respectively. For example, if the predicted image distortion index is 0.3, which is less than the preset image acquisition quality evaluation threshold, it is determined that the image acquisition is effective. This method realizes: obtaining the pressing force, placement angle, and acquisition time of the fingerprint image to be compared; processing the pressing force, placement angle, and acquisition time respectively in combination with corresponding preset parameter data to obtain a pressing force deviation rate, a placement angle deviation rate, and an acquisition time deviation rate; processing according to the pressing force deviation rate, angle deviation rate, and acquisition time deviation rate in combination with corresponding preset pressure weight values to obtain a predicted image distortion index, and comparing it with the preset image acquisition quality evaluation threshold to obtain the effectiveness of image acquisition.

[0035] It is worth mentioning that according to an embodiment of the present invention, it further includes: Performing block processing on the standard fingerprint image through a preset block method to obtain block fingerprint images; Obtaining the principal direction angle of the block fingerprint image and the direction angle of the minutiae points; Processing according to the principal direction angle and the direction angle to obtain an angle deviation; Comparing the angle deviation with a preset angle deviation threshold; If it is greater than the preset angle deviation threshold, then eliminate this minutiae point.

[0036] It should be noted that in order to accurately determine whether a minutiae point is noise interference, first perform block processing on the standard fingerprint image to obtain block fingerprint images. Those skilled in the art calculate the horizontal gradient and vertical gradient of each pixel point in the block fingerprint image, statistically calculate the gradients of all pixels in the block fingerprint image, calculate the average gradient direction to obtain the principal direction angle, and the direction angle of the minutiae point is calculated by those skilled in the art by tracing 3 - 5 pixels along the ridge line extension direction where the minutiae point is located and fitting the slope of the straight line; processing according to the principal direction angle and the direction angle to obtain an angle deviation, where the angle deviation is the absolute value of the difference between the principal direction angle and the direction angle. Finally, compare the obtained angle deviation with the preset angle deviation threshold. In this embodiment, the preset angle deviation threshold is set to 15°. If it is greater than 15°, it is determined that the minutiae point is noise interference and is given an elimination process. If it is less than or equal to 15°, then retain this minutiae point.

[0037] It is worth mentioning that according to an embodiment of the present invention, it further includes: Obtaining the density feature data, topological feature vector, and morphological feature data of the pores in the standard fingerprint image; Compare the density feature data, topological feature vector, and morphological feature data with the fingerprint feature data in the preset fingerprint database respectively to obtain a density similarity score, a topological feature similarity score, and a morphological similarity score; Process the density similarity score, topological feature similarity score, and morphological similarity score in combination with the preset density weight value, preset topological weight value, and preset morphological weight value to obtain a pore matching evaluation index; Compare the pore matching evaluation index with the preset pore matching quality evaluation threshold; If it is less than or equal to the preset pore matching quality evaluation threshold, it is determined that the pore comparison fails; If it is greater than the preset pore matching quality evaluation threshold, it is determined that the pore comparison is successful.

[0038] It should be noted that in order to improve the accuracy of fingerprint comparison, pore comparison is introduced. First, the fingerprint image is divided into 1mm×1mm grids, the number of pores in each grid is counted, a density distribution histogram is generated and normalized to a probability distribution to obtain the density feature data of the pores; by constructing a triangular mesh based on the pore center points, recording the length and angle of each side, and extracting the variance of the triangle interior angles and the mean side length to form a 128-dimensional feature vector to obtain the topological feature vector of the pores; by calculating the area, perimeter, aspect ratio, and Hu invariant moments for each individual pore to generate 7-dimensional features per pore to obtain the morphological feature data of the pores. Then, compare the similarity with the fingerprint feature data in the preset fingerprint database to obtain a density similarity score, a topological feature similarity score, and a morphological similarity score. Multiply the density similarity score by the preset density weight value, the topological feature similarity score by the preset topological weight value, and the morphological similarity score by the preset morphological weight value, and then sum them up to obtain the pore matching evaluation index. Finally, through threshold comparison, determine whether the pore comparison is successful to improve the anti-interference ability of fingerprint comparison; this method realizes: obtaining the density feature data, topological feature vector, and morphological feature data of the pores in the standard fingerprint image, and respectively comparing the similarity with the fingerprint feature data in the preset fingerprint database to obtain a density similarity score, a topological feature similarity score, and a morphological similarity score; processing the density similarity score, topological feature similarity score, and morphological similarity score in combination with the corresponding preset weight values to obtain a pore matching evaluation index; comparing the pore matching evaluation index with the preset pore matching quality evaluation threshold to obtain the pore comparison result.

[0039] The present invention also discloses a fingerprint comparison system based on multi-feature fusion, including a memory and a processor. The memory includes a fingerprint comparison method program based on multi-feature fusion. When the fingerprint comparison method program based on multi-feature fusion is executed by the processor, the following steps are implemented: Obtain the fingerprint image to be compared and perform preprocessing to obtain a standard fingerprint image; Extract according to the standard fingerprint image to obtain image quality evaluation data, process it to obtain an image quality evaluation index, and compare it with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint image; If the qualified status of the standard fingerprint image is qualified, extract features according to the standard fingerprint image to obtain global feature data and local feature data; Compare the global feature data and the local feature data with the fingerprint feature data in the preset fingerprint database respectively to obtain corresponding similarity scores, and perform weighted average processing to obtain a comparison similarity evaluation index; Compare the comparison similarity evaluation index with a preset comparison similarity evaluation threshold to obtain a comparison result.

[0040] It should be noted that in order to perform fingerprint comparison accurately and efficiently, first perform grayscale conversion, filter to remove noise, binarization, and thinning preprocessing on the collected fingerprint image to be compared to obtain a standard fingerprint image. To ensure that the quality of the collected image meets the standard, extract image quality evaluation data for processing to obtain an image quality evaluation index, and then perform threshold comparison to obtain the qualified status of the standard fingerprint image; if it is unqualified, output a warning reminder to prompt re-collection. If it is qualified, further perform global feature extraction and local feature extraction to provide data support for multi-feature fusion. Then compare the extracted global feature data and local feature data with the fingerprint feature data in the preset fingerprint database respectively to obtain corresponding similarity scores, and obtain a comparison similarity evaluation index through weighted average processing. Finally, determine whether the fingerprint comparison is successful through threshold comparison.

[0041] According to an embodiment of the present invention, the extracting according to the standard fingerprint image to obtain image quality evaluation data, processing it to obtain an image quality evaluation index, and comparing it with a preset image quality evaluation threshold to obtain the qualified status of the standard fingerprint image includes: Extract the standard fingerprint image to obtain image quality evaluation data, including image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data; Perform weighted summation processing on the image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity data to obtain an image quality evaluation index; Compare the image quality evaluation index with a preset image quality evaluation threshold; If it is less than or equal to the preset image quality evaluation threshold, determine that the standard fingerprint image is unqualified; If it is greater than the preset image quality evaluation threshold, determine that the standard fingerprint image is qualified.

[0042] It should be noted that before multi-feature extraction, the collected fingerprint images should be evaluated to prevent low-quality fingerprint images from affecting feature extraction and fingerprint comparison results. By extracting data such as image sharpness, image contrast, image signal-to-noise ratio, and ridge integrity, and performing weighted summation processing, an image quality evaluation index is obtained. Among them, the ridge integrity data is obtained by calculating the continuous length and the number of breaks of the ridges in the fingerprint image, and the corresponding weight values are obtained by querying a preset fingerprint comparison platform, and then threshold comparison is performed. According to the threshold comparison result, it is determined whether the standard fingerprint image is qualified.

[0043] According to an embodiment of the present invention, if the qualified state of the standard fingerprint image is qualified, then feature extraction is performed according to the standard fingerprint image to obtain global feature data and local feature data, including: If the qualified state of the standard fingerprint image is qualified, then the standard fingerprint image is subjected to feature extraction to obtain global feature data and local feature data; The global feature data includes orientation field feature data and frequency feature data; The local feature data includes minutiae feature data and local texture feature data.

[0044] It should be noted that in order to achieve multi - feature fusion processing, first, feature extraction is performed on the standard fingerprint image to obtain global feature data including orientation field feature data and frequency feature data. Among them, for the orientation field feature data, the Sobel operator is used on the fingerprint image to calculate the gradients in the x and y directions, obtaining the gradient magnitude and direction of each pixel point in the image. Then, the fingerprint image is divided into several appropriately sized blocks, usually blocks of 16×16 or 32×32 pixels. For each block, the gradient direction histogram of the pixel points therein is statistically analyzed to represent the dominant direction of the fingerprint ridges in the block, thereby obtaining the orientation field of the fingerprint. For the frequency feature data, the fingerprint image is filtered using a Gabor filter, and the frequency and direction parameters of the Gabor filter are matched with the frequency and direction of the fingerprint ridges. Through filtering operations at different frequencies and directions, the responses of the fingerprint image at different frequencies are obtained, and then statistical analysis is performed on the filtered image, such as calculating the energy distribution in each frequency channel as the frequency feature of the fingerprint; local feature data including minutiae feature data and local texture feature data. Among them, for the minutiae feature data, the fingerprint image is binarized to convert it into a black - and - white image to highlight the difference between the fingerprint ridges and the background. Then, the binarized image is thinned to obtain a single - pixel - wide fingerprint ridge skeleton. Finally, on the thinned image, by analyzing the endpoint and bifurcation point structures of the ridges, the minutiae of the fingerprint are detected. The local texture feature data is obtained by dividing the fingerprint image into smaller local regions, such as small blocks of 8×8 or 16×16 pixels, and then for each small block, its texture features are calculated, such as gray - level co - occurrence matrix (GLCM) features. The GLCM can describe the spatial relationship between different gray - level values in the image, and contrast, correlation, energy, and entropy are extracted from it to characterize the characteristics of the local texture of the fingerprint.

[0045] According to an embodiment of the present invention, it further includes: Obtain the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period; Perform weighted summation processing on the basis of the accuracy rate, false recognition rate, rejection rate, and equal error rate in combination with corresponding preset weight values to obtain an accuracy evaluation index for fingerprint comparison; Process the accuracy evaluation index and a preset accuracy requirement index to obtain a comparison accuracy deviation rate; Compare the comparison accuracy deviation rate with a preset comparison accuracy level threshold, and obtain an accuracy evaluation level according to the threshold range into which the comparison accuracy deviation rate falls; Query a preset feature weight value list according to the accuracy evaluation level to obtain dynamic feature weight values, including a first feature weight value for global feature data and a second feature weight value for local feature data.

[0046] It should be noted that when performing feature fusion, if a fixed feature weight value is used, the adaptability is insufficient. Therefore, in this embodiment, a dynamic feature weight value is adopted. To obtain the dynamic feature weight value, first, the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period are obtained. Among them, the accuracy rate is the ratio of the number of correctly matched fingerprint pairs to the total number of comparison pairs. The false recognition rate is the probability of mis-matching the fingerprints of different people as those of the same person. The rejection rate is the probability of mis-determining the fingerprints of the same person as those of different people. The equal error rate is the error rate when the false recognition rate and the rejection rate are equal. Then, a weighted summation process is performed in combination with the corresponding preset weight values to obtain an accuracy evaluation index for fingerprint comparison. Then, the accuracy evaluation index is processed with a preset accuracy requirement index to obtain a comparison accuracy deviation rate. The comparison accuracy deviation rate is the ratio of the absolute value of the difference between the accuracy evaluation index and the preset accuracy requirement index to the preset accuracy requirement index. Then, the comparison accuracy deviation rate is compared with a preset comparison accuracy level threshold. In this embodiment, the preset comparison accuracy level threshold is set to (0, 0.35], (0.35, 0.55], (0.55, 1], corresponding to first-level accuracy, second-level accuracy, and third-level accuracy respectively, where the first-level accuracy level is the highest. For example, if the obtained accuracy evaluation index is 7 and the preset accuracy requirement index is 10, then 3 / 10 = 0.3 is the comparison accuracy deviation rate, and the accuracy evaluation level is determined to be first-level accuracy. Finally, by querying a preset feature weight value list, a first feature weight value for global feature data and a second feature weight value for local feature data are obtained. Among them, the preset feature weight value list is obtained by querying a preset fingerprint comparison platform. This method realizes: obtaining the accuracy rate, false recognition rate, rejection rate, and equal error rate of fingerprint comparison within a preset time period, and performing a weighted summation process to obtain an accuracy evaluation index for fingerprint comparison; processing the accuracy evaluation index with a preset accuracy requirement index to obtain a comparison accuracy deviation rate, comparing it with a preset comparison accuracy level threshold, and obtaining an accuracy evaluation level according to the threshold range into which the comparison accuracy deviation rate falls; querying a preset feature weight value list according to the accuracy evaluation level to obtain a dynamic feature weight value.

[0047] According to an embodiment of the present invention, comparing the global feature data and the local feature data with the fingerprint feature data in a preset fingerprint database respectively to obtain corresponding similarity scores, and performing a weighted average process to obtain a comparison similarity evaluation index, including: Comparing the global feature data and the local feature data with the fingerprint feature data in a preset fingerprint database respectively to obtain corresponding similarity scores, including a global feature similarity score and a local feature similarity score; Performing a weighted average process on the global feature similarity score and the local feature similarity score in combination with the first feature weight value and the second feature weight value to obtain a comparison similarity evaluation index.

[0048] It should be noted that when performing feature fusion, first, the global feature data and the local feature data are respectively compared with the fingerprint feature data in the preset fingerprint database to obtain corresponding similarity scores, and then weighted average processing is performed by combining the first feature weight value and the second feature weight value to obtain a comparison similarity evaluation index; The calculation formula of the comparison similarity evaluation index is: ; Wherein, is the comparison similarity evaluation index, and are the global feature similarity score and the local feature similarity score respectively, and are the first feature weight value and the second feature weight value respectively.

[0049] According to the embodiment of the present invention, comparing the comparison similarity evaluation index with a preset comparison similarity evaluation threshold to obtain a comparison result includes: Comparing the comparison similarity evaluation index with a preset comparison similarity requirement index to obtain a comparison similarity relative value; Comparing the comparison similarity relative value with a preset comparison similarity evaluation threshold, wherein the preset comparison similarity evaluation threshold includes a first preset comparison similarity evaluation threshold and a second preset comparison similarity evaluation threshold; If it is less than or equal to the first preset comparison similarity evaluation threshold, it is determined that the comparison result is a comparison failure; If it is greater than the first preset comparison similarity evaluation threshold and less than or equal to the second preset comparison similarity evaluation threshold, it is determined that the comparison result is pending confirmation; If it is greater than the second preset comparison similarity evaluation threshold, it is determined that the comparison result is a comparison success.

[0050] It should be noted that in order to determine the fingerprint comparison result, first, the obtained comparison similarity evaluation index is compared with the preset comparison similarity requirement index to obtain the relative value of the comparison similarity. For example, if the comparison similarity evaluation index is 8 and the preset comparison similarity requirement index is 10, then 8 / 10 = 0.8 is the relative value of the comparison similarity. Then, it is respectively compared with the first preset comparison similarity evaluation threshold and the second preset comparison similarity evaluation threshold to obtain a comparison result including comparison failure, comparison result to be confirmed, or comparison success. In this embodiment, the preset comparison similarity evaluation thresholds are set as (0, 0.7], (0.7, 0.8], and (0.8, 1], corresponding to comparison failure, comparison result to be confirmed, and comparison success respectively. For example, if the obtained relative value of the comparison similarity is 0.75, then the comparison result is determined to be a comparison result to be confirmed.

[0051] According to an embodiment of the present invention, it further includes: Obtain the acquisition distortion evaluation data of the fingerprint image to be compared, including the pressing force, placement angle, and acquisition time; Process the pressing force with the preset minimum pressing force and the preset maximum pressing force to obtain the pressing force deviation rate; Compare the placement angle with the preset placement angle parameter data to obtain the placement angle deviation rate; Compare the acquisition time with the preset acquisition time parameter data to obtain the acquisition time deviation rate; Process according to the pressing force deviation rate, placement angle deviation rate, and acquisition time deviation rate in combination with the preset pressure weight value, preset placement angle weight value, and preset acquisition time weight value to obtain the predicted image distortion index; Compare the predicted image distortion index with the preset image acquisition quality evaluation threshold; If the predicted image distortion index is less than or equal to the preset image acquisition quality evaluation threshold, it is determined that the image acquisition is effective; If the predicted image distortion index is greater than the preset image acquisition quality evaluation threshold, it is determined that the image acquisition is invalid.

[0052] It should be noted that during fingerprint acquisition, excessive pressing force may distort the fingerprint ridges, insufficient force may result in an incomplete fingerprint image, incorrect placement angle may cause fingerprint image distortion, too short acquisition time may fail to obtain a complete and clear fingerprint, and too long acquisition time may cause blurring due to finger movement. To evaluate the quality of fingerprint image acquisition, acquisition distortion evaluation data including pressing force, placement angle, and acquisition time are obtained, and processed in combination with corresponding preset parameter data to obtain a pressing force deviation rate, a placement angle deviation rate, and an acquisition time deviation rate. Among them, the pressing force deviation rate is the ratio of the difference between the pressing force and the minimum preset pressing force to the difference between the maximum preset pressing force and the minimum preset pressing force. If the obtained pressing force is less than or equal to the minimum pressing force, a warning is directly output. For example, if the pressing force is 7, the minimum preset pressing force is 5, and the maximum preset pressing force is 10, then (7 - 5) / (10 - 5)=0.4 is the pressing force deviation rate. The placement angle deviation rate is the ratio of the absolute value of the difference between the placement angle and the preset placement angle parameter data to the preset placement angle parameter data, and the acquisition time deviation rate is the ratio of the absolute value of the difference between the acquisition time and the preset acquisition time parameter data to the preset acquisition time parameter data. According to the obtained pressing force deviation rate, placement angle deviation rate, and acquisition time deviation rate, combined with the preset pressure weight value, preset placement angle weight value, and preset acquisition time weight value, a predicted image distortion index is obtained. The calculation formula for the predicted image distortion index is: ; Among them, is the predicted image distortion index, 、 、 are the pressing force deviation rate, the placement angle deviation rate, and the acquisition time deviation rate respectively, 、 、 are the preset pressure weight value, the preset placement angle weight value, and the preset acquisition time weight value respectively, is a preset feature coefficient (the feature coefficient is obtained by querying through a preset fingerprint comparison platform); finally, compare the obtained predicted image distortion index with the preset image acquisition quality evaluation threshold. In this embodiment, the preset image acquisition quality evaluation threshold is set to (0, 0.55] and (0.55, 1], corresponding to effective image acquisition and ineffective image acquisition respectively. For example, if the predicted image distortion index is 0.3, which is less than the preset image acquisition quality evaluation threshold, it is determined that the image acquisition is effective. This method realizes: obtaining the pressing force, placement angle, and acquisition time of the fingerprint image to be compared; processing the pressing force, placement angle, and acquisition time respectively in combination with the corresponding preset parameter data to obtain the pressing force deviation rate, placement angle deviation rate, and acquisition time deviation rate; processing according to the pressing force deviation rate, angle deviation rate, and acquisition time deviation rate in combination with the corresponding preset pressure weight value to obtain the predicted image distortion index, and comparing it with the preset image acquisition quality evaluation threshold to obtain the effectiveness of image acquisition.

[0053] It is worth mentioning that according to an embodiment of the present invention, it further includes: Performing block processing on the standard fingerprint image through a preset block method to obtain block fingerprint images; Obtaining the main direction angle of the block fingerprint image and the direction angle of the minutiae points; Processing according to the main direction angle and the direction angle to obtain an angle deviation; Comparing the angle deviation with a preset angle deviation threshold; If it is greater than the preset angle deviation threshold, then eliminate this minutiae point.

[0054] It should be noted that in order to accurately determine whether a minutiae point is noise interference, first perform block processing on the standard fingerprint image to obtain block fingerprint images. Those skilled in the art calculate the horizontal gradient and vertical gradient of each pixel point in the block fingerprint image, statistically calculate the gradients of all pixels in the block fingerprint image, calculate the average gradient direction to obtain the main direction angle, and the direction angle of the minutiae point is obtained by those skilled in the art by tracing 3 - 5 pixels along the ridge line extension direction where the minutiae point is located, fitting the straight line slope, and calculating. Processing according to the main direction angle and the direction angle to obtain an angle deviation. The angle deviation refers to the absolute value of the difference between the main direction angle and the direction angle. Finally, compare the obtained angle deviation with the preset angle deviation threshold. In this embodiment, the preset angle deviation threshold is set to 15°. If it is greater than 15°, it is determined that the minutiae point is noise interference and is given an elimination process. If it is less than or equal to 15°, then retain this minutiae point.

[0055] It is worth mentioning that according to an embodiment of the present invention, it further includes: Obtaining the density feature data, topological feature vector, and morphological feature data of the sweat pores in the standard fingerprint image; The density feature data, topological feature vector, and morphological feature data are respectively compared with the fingerprint feature data in the preset fingerprint database to obtain a density similarity score, a topological feature similarity score, and a morphological similarity score; The density similarity score, topological feature similarity score, and morphological similarity score are processed in combination with a preset density weight value, a preset topological weight value, and a preset morphological weight value to obtain a pore matching evaluation index; The pore matching evaluation index is compared with a preset pore matching quality evaluation threshold; If it is less than or equal to the preset pore matching quality evaluation threshold, it is determined that the pore comparison fails; If it is greater than the preset pore matching quality evaluation threshold, it is determined that the pore comparison is successful.

[0056] It should be noted that in order to improve the accuracy of fingerprint comparison, pore comparison is introduced. First, the fingerprint image is divided into 1mm×1mm grids, the number of pores in each grid is counted, a density distribution histogram is generated, and it is normalized to a probability distribution to obtain the density feature data of the pores; by constructing a triangular mesh based on the center points of the pores, recording the length and angle of each side, and extracting the variance of the triangle interior angles and the average side length, a 128-dimensional feature vector is formed to obtain the topological feature vector of the pores; by calculating the area, perimeter, aspect ratio, and Hu invariant moments for each individual pore, 7-dimensional features per pore are generated to obtain the morphological feature data of the pores. Then, it is compared with the fingerprint feature data in the preset fingerprint database to obtain a density similarity score, a topological feature similarity score, and a morphological similarity score. The density similarity score is multiplied by the preset density weight value, the topological feature similarity score is multiplied by the preset topological weight value, the morphological similarity score is multiplied by the preset morphological weight value, and then the sum is processed to obtain the pore matching evaluation index. Finally, through threshold comparison, it is determined whether the pore comparison is successful, improving the anti-interference ability of fingerprint comparison; this method realizes: obtaining the density feature data, topological feature vector, and morphological feature data of the pores in the standard fingerprint image, and respectively comparing them with the fingerprint feature data in the preset fingerprint database to obtain a density similarity score, a topological feature similarity score, and a morphological similarity score; processing the density similarity score, topological feature similarity score, and morphological similarity score in combination with the corresponding preset weight values to obtain a pore matching evaluation index; comparing the pore matching evaluation index with the preset pore matching quality evaluation threshold to obtain the pore comparison result.

[0057] A fingerprint comparison method and system based on multi-feature fusion disclosed by the present invention realizes improving the accuracy and reliability of fingerprint comparison based on multi-feature fusion by evaluating a standard fingerprint image, multi-feature extraction, and dynamic weight fusion processing.

[0058] In several embodiments provided by this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the couplings, direct couplings, or communication connections between the various components shown or discussed can be through some interfaces. The indirect couplings or communication connections of devices or units can be electrical, mechanical, or other forms.

[0059] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0060] In addition, each functional unit in the embodiments of the present invention can be all integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in a unit. The above-mentioned integrated units can be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0061] Those of ordinary skill in the art can understand that all or part of the steps for implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a readable storage medium. When the program is executed, it executes the steps including the above method embodiments. The foregoing storage medium includes: removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks and other various media that can store program codes.

[0062] Alternatively, if the above-mentioned integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: removable storage devices, ROM, RAM, magnetic disks, or optical disks and other various media that can store program codes.

Claims

1. A fingerprint comparison method based on multi-feature fusion, characterized in that: The following steps are involved: Obtaining a fingerprint image to be compared, and performing preprocessing to obtain a standard fingerprint image; Extracting according to the standard fingerprint image to obtain image quality evaluation data, and processing to obtain an image quality evaluation index, and performing a threshold comparison with a preset image quality evaluation threshold to obtain a qualified state of the standard fingerprint image; If the qualified state of the standard fingerprint image is qualified, feature extraction is performed according to the standard fingerprint image to obtain global feature data and local feature data; The global feature data and the local feature data are respectively compared with the fingerprint feature data in the preset fingerprint library to obtain corresponding similarity scores, and weighted average processing is performed to obtain a comparison similarity evaluation index; The comparison similarity evaluation index is compared with a preset comparison similarity evaluation threshold to obtain a comparison result.

2. The fingerprint comparison method based on multi-feature fusion according to claim 1 is characterized in that: The step of extracting the standard fingerprint image to obtain image quality evaluation data, processing the image quality evaluation index, and performing a threshold comparison with a preset image quality evaluation threshold to obtain a qualified state of the standard fingerprint image includes: Extracting the standard fingerprint image to obtain image quality evaluation data, including image clarity, image contrast, image signal-to-noise ratio and ridge integrity data; Performing weighted sum processing on the image clarity, image contrast, image signal-to-noise ratio and grain integrity data to obtain an image quality evaluation index; Performing a threshold comparison between the image quality evaluation index and a preset image quality evaluation threshold; If it is less than or equal to the preset image quality assessment threshold, the standard fingerprint image is judged to be unqualified; If it is greater than the preset image quality assessment threshold, the standard fingerprint image is determined to be qualified.

3. The fingerprint comparison method based on multi-feature fusion according to claim 2 is characterized in that: If the qualified state of the standard fingerprint image is qualified, feature extraction is performed according to the standard fingerprint image to obtain global feature data and local feature data, including: If the qualified state of the standard fingerprint image is qualified, extracting features from the standard fingerprint image to obtain global feature data and local feature data; The global feature data includes direction field feature data and frequency feature data; The local feature data includes detail point feature data and local texture feature data.

4. The fingerprint comparison method based on multi-feature fusion according to claim 3 is characterized in that: Also includes: Obtain the accuracy rate, false recognition rate, rejection rate and error rate of fingerprint comparison within a preset time period; A weighted summation process is performed according to the accuracy rate, false recognition rate, rejection rate and equal error rate in combination with corresponding preset weight values ​​to obtain an accuracy evaluation index of the fingerprint comparison; Processing the accuracy assessment index with a preset accuracy requirement index to obtain a comparison accuracy deviation rate; Comparing the comparison accuracy deviation rate with a preset comparison accuracy level threshold, and obtaining an accuracy assessment level according to the threshold range within which the comparison accuracy deviation rate falls; A preset feature weight value list is queried according to the accuracy evaluation level to obtain dynamic feature weight values, including a first feature weight value of the global feature data and a second feature weight value of the local feature data.

5. The fingerprint comparison method based on multi-feature fusion according to claim 4 is characterized in that: The global feature data and the local feature data are respectively compared with the fingerprint feature data in the preset fingerprint library to obtain the corresponding similarity scores, and weighted average processing is performed to obtain the comparison similarity evaluation index, including: Comparing the global feature data and the local feature data with the fingerprint feature data in a preset fingerprint library respectively to obtain corresponding similarity scores, including a global feature similarity score and a local feature similarity score; The global feature similarity score and the local feature similarity score are combined with the first feature weight value and the second feature weight value to perform weighted averaging processing to obtain a comparison similarity evaluation index.

6. The fingerprint comparison method based on multi-feature fusion according to claim 5 is characterized in that: The comparing the comparison similarity evaluation index with a preset comparison similarity evaluation threshold to obtain a comparison result includes: Comparing the comparison similarity evaluation index with a preset comparison similarity requirement index to obtain a comparison similarity relative value; Performing a threshold comparison between the comparison similarity relative value and a preset comparison similarity evaluation threshold, wherein the preset comparison similarity evaluation threshold includes a first preset comparison similarity evaluation threshold and a second preset comparison similarity evaluation threshold; If it is less than or equal to the first preset comparison similarity evaluation threshold, the comparison result is determined to be a comparison failure; If it is greater than the first preset comparison similarity evaluation threshold and less than or equal to the second preset comparison similarity evaluation threshold, the comparison result is determined to be a comparison result to be confirmed; If it is greater than the second preset comparison similarity evaluation threshold, the comparison result is determined to be a successful comparison.

7. The fingerprint comparison method based on multi-feature fusion according to claim 6 is characterized in that: Also includes: Obtaining acquisition distortion evaluation data of the fingerprint image to be compared, including pressing pressure, placement angle and acquisition time; Processing the pressing force with a preset minimum pressing force value and a preset maximum pressing force value to obtain a pressing force deviation rate; Comparing the placement angle with preset placement angle parameter data to obtain a placement angle deviation rate; Comparing the acquisition time with preset acquisition time parameter data to obtain an acquisition time deviation rate; The predicted image distortion index is obtained by processing the pressure deviation rate, the placement angle deviation rate and the acquisition time deviation rate in combination with a preset pressure weight value, a preset placement angle weight value and a preset acquisition time weight value; Performing a threshold comparison between the predicted image distortion index and a preset image acquisition quality assessment threshold; If the predicted image distortion index is less than or equal to the preset image acquisition quality assessment threshold, the image acquisition is determined to be valid; If the predicted image distortion index is greater than a preset image acquisition quality assessment threshold, the image acquisition is determined to be invalid.

8. A fingerprint matching system based on multi-feature fusion, characterized in that: The invention comprises a memory and a processor, wherein the memory comprises a program of a fingerprint comparison method based on multi-feature fusion, and the fingerprint comparison method program based on multi-feature fusion is executed by the processor to implement the following steps: Obtaining a fingerprint image to be compared, and performing preprocessing to obtain a standard fingerprint image; Extracting according to the standard fingerprint image to obtain image quality evaluation data, and processing to obtain an image quality evaluation index, and performing a threshold comparison with a preset image quality evaluation threshold to obtain a qualified state of the standard fingerprint image; If the qualified state of the standard fingerprint image is qualified, feature extraction is performed according to the standard fingerprint image to obtain global feature data and local feature data; The global feature data and the local feature data are respectively compared with the fingerprint feature data in the preset fingerprint library to obtain corresponding similarity scores, and weighted average processing is performed to obtain a comparison similarity evaluation index; The comparison similarity evaluation index is compared with a preset comparison similarity evaluation threshold to obtain a comparison result.

9. The fingerprint matching system based on multi-feature fusion according to claim 8, characterized in that: The step of extracting the standard fingerprint image to obtain image quality evaluation data, processing the image quality evaluation index, and performing a threshold comparison with a preset image quality evaluation threshold to obtain a qualified state of the standard fingerprint image includes: Extracting the standard fingerprint image to obtain image quality evaluation data, including image clarity, image contrast, image signal-to-noise ratio and ridge integrity data; Performing weighted sum processing on the image clarity, image contrast, image signal-to-noise ratio and grain integrity data to obtain an image quality evaluation index; Performing a threshold comparison between the image quality evaluation index and a preset image quality evaluation threshold; If it is less than or equal to the preset image quality assessment threshold, the standard fingerprint image is judged to be unqualified; If it is greater than the preset image quality assessment threshold, the standard fingerprint image is determined to be qualified.

10. The fingerprint matching system based on multi-feature fusion according to claim 9, characterized in that: If the qualified state of the standard fingerprint image is qualified, feature extraction is performed according to the standard fingerprint image to obtain global feature data and local feature data, including: If the qualified state of the standard fingerprint image is qualified, extracting features from the standard fingerprint image to obtain global feature data and local feature data; The global feature data includes direction field feature data and frequency feature data; The local feature data includes detail point feature data and local texture feature data.

Images