A video-based dynamic palmprint ROI recognition method and system

Abstract

The present application relates to the technical field of image processing and analysis, and particularly relates to a dynamic palmprint ROI recognition method and system based on video, which comprises the following steps: acquiring a video frame containing a user's palm; selecting a video frame containing a palm and having a palm boundary as an effective frame; selecting an effective frame meeting image condition requirements as a key frame; extracting a palm ROI and palmprint features; registering the palmprint; matching the user's palmprint features with registered palmprints to perform palmprint recognition. The present application can improve the flexibility of image shooting by acquiring a palm video image, screening key frames therefrom, effectively avoiding problems such as focus blur caused by movement, improving the efficiency of palmprint collection by frame-by-frame screening, obtaining more abundant information from dynamic video data or continuous multiple frames of images, being conducive to feature extraction and identification, and improving the convenience and practicality of palmprint recognition on the premise of retaining abundant image information.

Description

Technical Field

[0001] This invention relates to the field of image processing and analysis technology, and in particular to a method and system for identifying dynamic palmprint ROIs based on video. Background Technology

[0002] The information and networked society places higher demands on the security of various information and systems, and identity authentication is the core cornerstone for strengthening information and system security. Traditional identity authentication technologies include item-based and knowledge-based methods. Items include ID cards, passports, keys, etc., which are easily lost, damaged, stolen, and counterfeited, giving imposters or attackers opportunities. Knowledge includes passwords, PINs, verification codes, etc., which are easily forgotten and cracked by attackers. In contrast to these traditional identity verification methods, biometric identification technology extracts inherent physiological or behavioral characteristics of the human body, effectively overcoming the shortcomings of traditional identity authentication methods. Currently, various biometrics are used for identity authentication, such as fingerprints, faces, irises, and palm prints; among them, palm print recognition, as an important supplement to existing biometric technologies, has the characteristics of rich identification information, low resolution requirements, uniqueness, and low cost.

[0003] Currently, most palmprint recognition methods process and analyze single static images, which significantly limits the amount of information obtained from palmprint features. Furthermore, current palmprint acquisition methods have many drawbacks, including:

[0004] (1) Contact-based data collection poses health and personal safety risks. Contact-based data collection will undoubtedly increase the risk of contracting infectious diseases; and contact-based data collection will limit the user's freedom, flexibility and comfort.

[0005] (2) When users directly take palm print pictures, they are easily affected by complex backgrounds, changes in lighting and changes in hand posture. The image acquisition accuracy is not high, and there are also problems such as motion blur and focus blur. Moreover, for a single static image, the information contained is often limited, and if the user's palm is not positioned properly, the user needs to adjust the palm position multiple times to take pictures repeatedly.

[0006] (3) The performance of the shooting elements of different terminals varies, making it difficult to preprocess different images and extract palm print features from them;

[0007] The aforementioned shortcomings lead to difficulties in palmprint extraction and recognition. Existing methods struggle to extract effective and sufficient palmprint information flexibly from images and to completely remove interference from the external environment during the shooting and preprocessing process, which is detrimental to improving the efficiency of palmprint recognition. Summary of the Invention

[0008] This invention provides a video-based dynamic palmprint ROI recognition method and system to address the shortcomings of existing technologies that struggle to obtain effective palmprint information from a single palmprint image and involve excessively complex image preprocessing. By using video as a carrier, auxiliary palmprint feature acquisition is achieved on the terminal, which helps reduce the complexity of palmprint recognition preprocessing. Using video as a carrier also helps improve the stability and robustness of the recognition system.

[0009] This invention provides a video-based method for recognizing dynamic palmprint ROIs, comprising the following steps:

[0010] S1 acquires a video image of the user's palm, and acquires each video frame containing the user's palm;

[0011] S2 analyzes each video frame one by one, and selects video frames that contain a hand and have a hand boundary as valid frames.

[0012] S3 analyzes each of the valid frames one by one, obtains the image clarity and image highlight of the valid frames, removes valid frames whose image clarity and image highlight do not meet the preset conditions, and selects the valid frames that meet the preset conditions as key frames.

[0013] S4 obtains palm boundary and finger valley point location information from the preset region in the key frame, extracts palm ROI, and extracts palm print features;

[0014] S5 registers the palmprint based on the palmprint features; acquires a real-time video image of the user's palm, acquires the user's palmprint features based on steps S1-S4, and matches the user's palmprint features with the features of the registered palmprint to perform palmprint recognition.

[0015] Specifically, in step S2, selecting valid frames from the video frames includes:

[0016] Each video frame is analyzed one by one, and a preset fixed area of ​​the video frame is extracted to determine whether there are three gaps between the four fingers. If so, it is determined that the video frame contains a palm.

[0017] If a hand is detected, the video frame is further cropped into preset areas at the top and bottom edges to determine if a hand boundary exists. If a hand boundary exists in both the preset areas at the top and bottom edges, the video frame is considered a valid frame.

[0018] Furthermore, in step S2, after determining that there are three gaps between the four fingers, the process includes:

[0019] Identify the finger gap area within the preset fixed area;

[0020] The finger gap area is converted into a grayscale image and then binarized.

[0021] Morphological closing operation is performed on the binarized grayscale image, followed by horizontal integration.

[0022] Obtain the lengths of the gaps between the little finger and the ring finger (r1), the gaps between the ring finger and the middle finger (r2), and the gaps between the middle finger and the index finger (r3), respectively.

[0023] Specifically, it is determined whether the gap r2 between the ring finger and the middle finger is the minimum value among the three gaps. If it is the minimum value, the current frame is determined to be a valid frame.

[0024] Specifically, in step S3, obtaining the image clarity of the valid frames and discarding valid frames whose image clarity does not meet the preset conditions includes:

[0025] Obtaining the image sharpness of the valid frame includes:

[0026] Obtain the information entropy evaluation function and calculate the information entropy:

[0027]

[0028]

[0029] Where b is a parameter, g is the image grayscale value, G is the maximum value of the image grayscale value, k represents the defocused image sequence, and P k (g) represents the probability of grayscale value g appearing in the k-th frame image, n g This represents the number of pixels with gray level g in the image, where M and N represent the width and height of the image, respectively.

[0030] Obtain the information entropy of each valid frame, normalize all entropy values, and discard the corresponding valid frame if there is a frame with a value of 0.

[0031] Specifically, in step S3, obtaining the image highlight of the valid frames and discarding valid frames whose image highlight does not meet the preset conditions includes:

[0032] Get image brightness m:

[0033]

[0034] Get image saturation s:

[0035]

[0036] Generate a brightness and saturation histogram:

[0037]

[0038] Where r, g, and b represent the red, green, and blue channels of a pixel, respectively, and each coordinate (m, s) in the histogram represents the number of pixels in the m and s values ​​of the original color image;

[0039] Calculate the number of pixels in the MS histogram that exceed the brightness threshold and compare it with the total number of pixels in the valid frame. If the number exceeds the preset brightness threshold, it is determined that there is a highlight. Valid frames containing the highlight are discarded, and the remaining valid frames are used as keyframes.

[0040] Specifically, step S4 includes:

[0041] Identify the upper boundary line on the index finger side and the lower boundary line on the little finger side of the palm, including:

[0042] Extract a rectangular region of a preset size between the upper and lower boundary lines, perform boundary enhancement operations to obtain the energy map of the rectangular region;

[0043] The energy map of the rectangular region is binarized, and the boundary positions are obtained based on the horizontal integral. The ordinates of the upper and lower boundary lines are obtained respectively.

[0044] Obtain the location information of the finger valley point, including:

[0045] The detection area is any rectangular region located at a preset distance below the upper boundary line in the valid frame.

[0046] The detection area is edge-enhanced using a multi-directional Gabor filter to obtain an energy map of the finger gap region.

[0047] Based on the energy map of the finger gap area, remove the pixels in the 0 direction, perform vertical integration, and obtain the horizontal coordinate of the finger gap point.

[0048] Specifically, in step S4, the ROI is extracted and normalized based on the palm boundary and fingertip location information, including:

[0049] Based on the ordinate Y of the upper boundary line H The ordinate of the lower boundary line (Y) L And the x-coordinate of the valley point b Determine ROI:

[0050]

[0051] L=Y H -Y L ;

[0052] The ROI is resized to obtain a 128*128 ROI.

[0053] Specifically, in step S5, registering the palmprint based on the palmprint features includes:

[0054] Obtain the Hamming distance between each keyframe and the ROI of the previous adjacent keyframe, and determine whether the Hamming distance is less than a preset distance threshold. If it is less than the distance threshold, then the current keyframe is used as the registration keyframe.

[0055] If the total number of processed keyframes reaches a preset total threshold, and the total number of registered keyframes is lower than the number of registered keyframes, then the registration of the palmprint feature fails.

[0056] This invention also provides a video-based dynamic palmprint ROI recognition system, comprising:

[0057] The palm video capture module is used to acquire video images of the user's palm and to capture each video frame containing the user's palm.

[0058] The video frame processing module is used to analyze each video frame one by one and select video frames that contain a hand and have a hand boundary as valid frames.

[0059] The video frame processing module further analyzes each of the valid frames one by one, obtains the image clarity and image highlight of the valid frames, removes valid frames whose image clarity and image highlight do not meet the preset conditions, and selects the valid frames that meet the preset conditions as key frames.

[0060] The image preprocessing module is used to obtain palm boundary and finger valley point location information from the preset region in the key frame, extract palm ROI, and extract palm print features;

[0061] The palmprint recognition module is used to register palmprints based on the palmprint features. The palmprint recognition module acquires real-time video images of the user's palm, acquires the user's palmprint features based on steps S1-S4, and matches the user's palmprint features with the features of the registered palmprints to perform palmprint recognition.

[0062] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the video-based dynamic palmprint ROI recognition method as described above.

[0063] The present invention provides a video-based dynamic palmprint ROI recognition method and system, which has at least the following technical advantages compared with the prior art:

[0064] (1) By acquiring video images of the user's palm, key frames can be selected from the video images of the user's palm, which can improve the flexibility of image shooting. Furthermore, by using video, the impact of background, lighting changes, and palm posture changes during shooting can be reduced, which can effectively avoid problems such as focus blur caused by motion. By selecting frames one by one from the video, the efficiency of palm print acquisition can be improved.

[0065] (2) By obtaining the palm boundary and finger valley point position information from the preset area in the key frame, extracting the palm ROI and palm print features, it is possible to obtain richer information from dynamic video data or continuous multi-frame images, which is beneficial for feature extraction and identification and ensures the accuracy requirements for practical application.

[0066] (3) It can automatically register and recognize users' palm print information. Users do not need to take photos manually during registration and recognition. They only need to scan and take videos, which improves the convenience of palm print recognition while preserving rich image information. Attached Figure Description

[0067] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0068] Figure 1 This is a flowchart illustrating the video-based dynamic palmprint ROI recognition method provided by the present invention.

[0069] Figure 2 This is one of the video frame processing diagrams of the video-based dynamic palmprint ROI recognition method provided by the present invention;

[0070] Figure 3 This is the second video frame processing diagram of the video-based dynamic palmprint ROI recognition method provided by the present invention;

[0071] Figure 4 This is the third video frame processing diagram of the video-based dynamic palmprint ROI recognition method provided by the present invention;

[0072] Figure 5 This is the fourth video frame processing diagram of the video-based dynamic palmprint ROI recognition method provided by the present invention;

[0073] Figure 6 This is the fifth video frame processing diagram of the video-based dynamic palmprint ROI recognition method provided by the present invention;

[0074] Figure 7 This is a schematic diagram of the horizontal integral projection of the finger gap region in the video-based dynamic palmprint ROI recognition method provided by the present invention.

[0075] Figure 8 This is a schematic diagram of the composition of the video-based dynamic palmprint ROI recognition system provided by the present invention. Detailed Implementation

[0076] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0077] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or modules is not limited to the steps or modules listed, but may optionally include steps or modules not listed, or may optionally include other steps or modules inherent to such process, method, product, or apparatus.

[0078] It should be noted that the terms "first" and "second" used in this invention merely distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first" and "second" can be interchanged in a specific order or sequence where permissible. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those described or illustrated herein.

[0079] In one embodiment, such as Figure 1 As shown, the present invention provides a video-based dynamic palmprint ROI recognition method, comprising the following steps:

[0080] S1 acquires a video image of the user's palm, and acquires each video frame containing the user's palm;

[0081] S2 analyzes each video frame one by one, and selects video frames that contain a hand and have a hand boundary as valid frames.

[0082] S3 analyzes each of the valid frames one by one, obtains the image clarity and image highlight of the valid frames, removes valid frames whose image clarity and image highlight do not meet the preset conditions, and selects the valid frames that meet the preset conditions as key frames.

[0083] S4 obtains palm boundary and finger valley point location information from the preset region in the key frame, extracts palm ROI, and extracts palm print features;

[0084] S5 registers the palmprint based on the palmprint features; acquires a real-time video image of the user's palm, acquires the user's palmprint features based on steps S1-S4, and matches the user's palmprint features with the features of the registered palmprint to perform palmprint recognition.

[0085] Optionally, in step S1, when acquiring a video image of the user's palm, a first auxiliary line, a second auxiliary line, and an auxiliary point are displayed on the screen of the mobile terminal; the first auxiliary line and the second auxiliary line do not intersect; the auxiliary point is an endpoint of the first auxiliary line;

[0086] The relative positions of the first and second auxiliary lines are fixed, but one possible scenario is that the first and second auxiliary lines are parallel to each other. That is, regardless of how the first and second auxiliary lines are adjusted, their positional relationship always remains parallel.

[0087] Furthermore, the distance between the first and second auxiliary lines, as well as the lengths of the first and second auxiliary lines, are fixed and proportional to the screen size of the mobile terminal. Specifically, the length of the first auxiliary line is less than the length of the second auxiliary line.

[0088] When taking photos, the user should position their hand with four fingers together and the thumb naturally spread. The hand position and guide lines should be as follows: Figure 2 As shown, this allows the palm to be controlled in a way that is easy to select and facilitates subsequent steps such as video frame filtering and image processing, thereby improving the reliability of palm print information appearing in the video.

[0089] Specifically, in step S2, selecting valid frames from the video frames includes:

[0090] Each video frame is analyzed one by one, and a preset fixed area of ​​the video frame is extracted to determine whether there are three gaps between the four fingers. If so, it is determined that the video frame contains a palm.

[0091] If a hand is detected, the video frame is further cropped into a preset area at the top and bottom edges to determine if a hand boundary exists. If a hand boundary exists in both the preset areas at the top and bottom edges, the video frame is considered a valid frame.

[0092] Conversely, if there are no three gaps between the four fingers, the message indicates that there is no palm, and the video frame is invalid. The process continues to read the next video frame.

[0093] Furthermore, extracting such Figure 2The shadowed area shown is used to extract features from the preset fixed area. Based on the characteristics of the finger gap area, it is determined whether the video frame contains a palm and whether the palm is placed in a suitable position. If the palm is present and correctly placed, the next step is performed. If the palm is not present, feedback is given that there is no palm and the next frame image is read.

[0094] Furthermore, preset areas are extracted from the top and bottom edges of the video frame, such as... Figure 3 The regions ① and ② shown by the two rectangles at the upper and lower edges are used to extract features from the regions. Based on whether the regions contain palm print boundaries, it is determined whether the video frame meets the requirements for extracting a valid ROI. If the conditions are met, the next step is performed. If the conditions are not met, feedback is provided to adjust the palm position.

[0095] Since there are differences between the palm and the background, it is determined whether the region contains the palm boundary, and then whether the palm in the video frame meets the conditions for palmprint feature extraction.

[0096] Furthermore, Figure 3 The coordinates of A, B, C, and D are A(4*h / 15, 14*w / 15), B(9*h / 15, 14*w / 15), C(7*h / 15, w / 15), and D(9*h / 15, w / 15), respectively. L is the distance between line segments AB and CD, and d = round(L / 10). The width * height of the dashed rectangular regions ① and ② are both d * (d + y). D ), where y D Let D be the ordinate of point D. The left boundary of ① passes through point D, and the left boundary of ② passes through point A. This means that there is a hand boundary within regions ① and ②, with CD and AB being the upper and lower boundaries of the hand, respectively.

[0097] It should be noted that, ideally, the preset first and second auxiliary lines coincide with the upper and lower boundary lines of the palm.

[0098] Furthermore, in step S2, after determining that there are three gaps between the four fingers, the process includes:

[0099] Identify the finger gap area within the preset fixed area;

[0100] The finger gap area is converted into a grayscale image and then binarized.

[0101] Morphological closing operation is performed on the binarized grayscale image, followed by horizontal integration.

[0102] Obtain the lengths of the gaps between the little finger and the ring finger (r1), the gaps between the ring finger and the middle finger (r2), and the gaps between the middle finger and the index finger (r3), respectively.

[0103] Specifically, it is determined whether the gap r2 between the ring finger and the middle finger is the minimum value among the three gaps. If it is the minimum value, the current frame is determined to be a valid frame.

[0104] Optionally, the finger gap area can be located based on the upper and lower boundaries of the palm and the vertical distance between them, such as... Figure 4 As shown;

[0105] The finger gap area is further converted into a grayscale image; the finger gap and skin areas are marked with white and gray respectively; since the ambient lighting is uncontrollable, the binarization threshold is set to:

[0106] t = v min +(v max -v min )×α;

[0107] Among them, v min and v max These are the minimum and maximum gray values ​​for the finger gap area, respectively, with an optimal adjustment factor α = 0.45;

[0108] Based on the aforementioned threshold t, a binarization operation is performed, and the binarization result is as follows: Figure 5 As shown:

[0109]

[0110] Morphological closing operations are performed on the binarized result to eliminate the numerous connected regions in the finger gap region, such as... Figure 6 As shown, the smoothness of the finger gaps is improved after processing, which facilitates subsequent horizontal integration and improves computational efficiency; further horizontal integration yields... Figure 7 The distribution results shown are based on Figure 7 The distribution results shown clearly indicate the location of the finger gap area;

[0111] Specifically, in step S3, keyframes are filtered based on valid frames, including:

[0112] Obtaining the image sharpness of the valid frames and discarding valid frames whose image sharpness does not meet the preset conditions includes:

[0113] Obtaining the image sharpness of the valid frame includes:

[0114] Obtain the information entropy evaluation function and calculate the information entropy:

[0115]

[0116]

[0117] Where b is a parameter, g is the image grayscale value, G is the maximum value of the image grayscale value, k represents the defocused image sequence, and P k (g) represents the probability of grayscale value g appearing in the k-th frame image, n g This represents the number of pixels with gray level g in the image, where M and N represent the width and height of the image, respectively; preferably, the value of b is 2.

[0118] Obtain the information entropy of each valid frame, normalize all entropy values, and discard the corresponding valid frame if there is a frame with a value of 0.

[0119] Specifically, in step S3, obtaining the image highlight of the valid frames and discarding valid frames whose image highlight does not meet the preset conditions includes:

[0120] Get image brightness m:

[0121]

[0122] Get image saturation s:

[0123]

[0124] Generate a brightness and saturation histogram:

[0125]

[0126] Where r, g, and b represent the red, green, and blue channels of a pixel, respectively, and each coordinate (m, s) in the histogram represents the number of pixels in the m and s values ​​of the original color image;

[0127] Calculate the number of pixels in the MS histogram that exceed the brightness threshold and compare it with the total number of pixels in the valid frame. If the number exceeds the preset brightness threshold, it is determined that there is a highlight and the valid frame containing the highlight is discarded.

[0128] Obtain the number of pixels exceeding the brightness threshold in the MS histogram of any frame image, and then compare it with the total number of pixels in the entire image. If the number exceeds the preset threshold, it is determined that there is a highlight, and the frame image is discarded. The remaining valid frames are used as keyframes.

[0129] Specifically, step S4 includes:

[0130] Identify the upper boundary line on the index finger side and the lower boundary line on the little finger side of the palm, including:

[0131] Extract a rectangular region of a preset size between the upper and lower boundary lines, perform boundary enhancement operations to obtain the energy map of the rectangular region;

[0132] The energy map of the rectangular region is binarized, and the boundary positions are obtained based on the horizontal integral. The ordinates of the upper and lower boundary lines are obtained respectively.

[0133] Obtain the location information of the finger valley point, including:

[0134] The detection area is any rectangular region located at a preset distance below the upper boundary line in the valid frame.

[0135] The detection area is edge-enhanced using a multi-directional Gabor filter to obtain an energy map of the finger gap region.

[0136] Based on the energy map of the finger gap area, remove the pixels in the 0 direction, perform vertical integration, and obtain the horizontal coordinate of the finger gap point.

[0137] Specifically, in step S4, the ROI is extracted and normalized based on the palm boundary and fingertip location information, including:

[0138] Based on the ordinate Y of the upper boundary line H The ordinate of the lower boundary line (Y) L And the x-coordinate of the valley point b Determine ROI:

[0139]

[0140] L=Y H -Y L ;

[0141] The ROI is resized to obtain a 128*128 ROI; the obtained ROI area is as follows. Figure 8 As shown;

[0142] It should be noted that ROI stands for Region of Interest, which is the region of interest in a dynamic palm print.

[0143] This invention employs an auxiliary approach, using video as a medium to achieve palmprint recognition. By using a keyframe and effective frame selection strategy, it avoids the influence of complex backgrounds, significant lighting changes, and subtle changes in hand posture under non-contact conditions. At the same time, by detecting blurred frames and frames containing highlights, it avoids the influence of blurred frames caused by motion blur and focus blur, as well as the influence of video frames containing highlights caused by strong light illumination.

[0144] Because it uses video as the carrier for palmprint registration and recognition, this method offers a better user experience compared to recognition methods that use static images. The use of boundary line calibration and finger valley point localization improves the robustness and stability of the system, enabling more effective acquisition of dynamic palmprint regions of interest, which is conducive to the rapid and accurate extraction of palmprint features.

[0145] Specifically, in step S5, registering the palmprint based on the palmprint features includes:

[0146] Obtain the Hamming distance between each keyframe and the ROI of the previous adjacent keyframe, and determine whether the Hamming distance is less than a preset distance threshold. If it is less than the distance threshold, then the current keyframe is used as the registration keyframe.

[0147] If the total number of processed keyframes reaches a preset total threshold, and the total number of registered keyframes is lower than the number of registered keyframes, then the registration of the palmprint feature fails.

[0148] As an example, during the palmprint authentication stage, the system reads the authentication palmprint video and processes the palmprint video frame by frame. The system processes the authentication video frame by frame to obtain the ROI of the key frame. The feature template of the ROI of the current frame is compared with all registered templates one by one. If the Hamming distance is less than the preset threshold, the counter is incremented by one until all registered templates have been compared in the current valid frame.

[0149] If the currently authenticated video frame is not the last frame and the counter value is greater than or equal to 80% of the total number of registered frames, then "authentication successful" is returned, and the authentication process ends.

[0150] If the currently authenticated video frame is not the last frame and the counter value is less than 80% of the total number of registered frames, then set the counter value to zero and repeat the above steps.

[0151] If none of the authentication frames meet the requirements, the system returns "Authentication failed" and ends the authentication process; at this point, the user is prompted to re-record the hand video.

[0152] In one embodiment, such as Figure 8 As shown, the present invention also provides a video-based dynamic palmprint ROI recognition system. The recognition system described below corresponds to the recognition method described above. The system includes:

[0153] The palm video capture module is used to acquire video images of the user's palm and to capture each video frame containing the user's palm.

[0154] The video frame processing module is used to analyze each video frame one by one and select video frames that contain a hand and have a hand boundary as valid frames.

[0155] The video frame processing module further analyzes each of the valid frames one by one, obtains the image clarity and image highlight of the valid frames, removes valid frames whose image clarity and image highlight do not meet the preset conditions, and selects the valid frames that meet the preset conditions as key frames.

[0156] The image preprocessing module is used to obtain palm boundary and finger valley point location information from the preset region in the key frame, extract palm ROI, and extract palm print features;

[0157] The palmprint recognition module is used to register palmprints based on the palmprint features. The palmprint recognition module acquires real-time video images of the user's palm, acquires the user's palmprint features based on steps S1-S4, and matches the user's palmprint features with the features of the registered palmprints to perform palmprint recognition.

[0158] Furthermore, it also includes an information feedback module, which is used to determine whether the palm is recognized in the judgment of valid frames and key frames; if the palm is not recognized, feedback information is given to the user, prompting the user to repeatedly adjust the palm position.

[0159] Furthermore, the information feedback module is also used to provide feedback on the results of palmprint registration using the palmprint features, and to provide feedback on the results of palmprint recognition and matching based on the input palmprint video.

[0160] Furthermore, the above-described methods can be implemented as software functional units, and when sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0161] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer is able to execute the dynamic palmprint ROI recognition method provided by the above methods, the steps of which are consistent with the steps described in the above embodiments, and will not be repeated here.

[0162] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the dynamic palmprint ROI recognition method provided by the above methods. The steps of the method are consistent with the steps described in the above embodiments, and will not be repeated here.

[0163] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0164] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for recognizing dynamic palmprint ROIs based on video, characterized in that, include: S1 acquires a video image of the user's palm, and acquires each video frame containing the user's palm; S2 analyzes each video frame one by one, and selects video frames that contain a hand and have a hand boundary as valid frames. S3 analyzes each of the valid frames one by one, obtains the image clarity and image highlight of the valid frames, removes valid frames whose image clarity and image highlight do not meet the preset conditions, and selects the valid frames that meet the preset conditions as key frames. S4 obtains palm boundary and finger valley point location information from the preset region in the key frame, extracts palm ROI, and extracts palm print features; S5 registers the palmprint based on the palmprint features; acquires a real-time video image of the user's palm, acquires the user's palmprint features based on steps S1-S4, matches the user's palmprint features with the features of the registered palmprint, and performs palmprint recognition. Step S4 includes: Identifying the upper boundary line on the index finger side and the lower boundary line on the little finger side of the palm includes: extracting a rectangular region of a preset size between the upper and lower boundary lines, performing boundary enhancement operations to obtain an energy map of the rectangular region; performing a binarization operation on the energy map of the rectangular region, obtaining the boundary position based on horizontal integration, and obtaining the ordinates of the upper and lower boundary lines respectively. Obtaining the finger gap location information includes: obtaining any rectangular region at a preset distance below the upper boundary line in the effective frame as the detection region; performing edge enhancement on the detection region using a multi-directional Gabor filter to obtain the finger gap region energy map; and based on the finger gap region energy map, removing pixels in the 0 direction, performing vertical integration, and obtaining the horizontal coordinate of the finger gap point.

2. The method for identifying a dynamic palmprint ROI based on video according to claim 1, characterized in that, In step S2, selecting valid frames from the video frames includes: Each video frame is analyzed one by one, and a preset fixed area of ​​the video frame is extracted to determine whether there are three gaps between the four fingers. If so, it is determined that the video frame contains a palm. If a hand is detected, the video frame is further cropped into preset areas at the top and bottom edges to determine if a hand boundary exists. If a hand boundary exists in both the preset areas at the top and bottom edges, the video frame is considered a valid frame.

3. The method for identifying a dynamic palmprint ROI based on video according to claim 2, characterized in that, In step S2, after determining that there are three gaps between the four fingers, the following steps are taken: Identify the finger gap area within the preset fixed area; The finger gap area is converted into a grayscale image and then binarized. Morphological closing operation is performed on the binarized grayscale image, followed by horizontal integration. Obtain the length of the gap between the little finger and the ring finger respectively. r 1. The length of the space between the ring finger and the middle finger r 2. The length of the space between the middle and index fingers r 3; Determine the length of the gap between the ring finger and the middle finger r Is 2 the minimum value among the three gaps? If it is the minimum value, then the current frame is considered a valid frame.

4. A method for identifying a dynamic palmprint ROI based on video according to claim 1 or 3, characterized in that, In step S3, the image clarity of the valid frames is obtained, and valid frames whose image clarity does not meet the preset conditions are removed, including: Obtaining the image sharpness of the valid frame includes: Obtain the information entropy evaluation function and calculate the information entropy: ； ； in, For parameters, The image grayscale value, The maximum value of the image's grayscale. Represents a sequence of out-of-focus images. In the first grayscale values ​​in a frame image The probability of occurrence This indicates that the gray level in the image is... The number of pixels that appear. and These represent the width and height of the image, respectively. Obtain the information entropy of each valid frame, normalize all entropy values, and discard the corresponding valid frame if there is a frame with a value of 0.

5. A method for identifying a dynamic palmprint ROI based on video according to claim 1 or 3, characterized in that, In step S3, the image highlight of the valid frames is obtained, and valid frames whose image highlight does not meet the preset conditions are removed, including: Get image brightness : ； Get image saturation : ； Generate a brightness and saturation histogram: ； in 、 、 The red, green, and blue channels of a pixel are represented by their respective coordinates in the histogram. Represents the original color image. Value and The number of pixels in the value; Calculate the number of pixels in the MS histogram that exceed the brightness threshold and compare it with the total number of pixels in the valid frame. If the number exceeds the preset brightness threshold, it is determined that there is a highlight. Valid frames containing the highlight are discarded, and the remaining valid frames are used as keyframes.

6. The method for identifying a dynamic palmprint ROI based on video according to claim 1, characterized in that, In step S4, the ROI is extracted and normalized based on the palm boundary and fingertip location information, including: Based on the ordinate of the upper boundary line , the ordinate of the lower boundary line and the x-coordinate of the valley point Determine ROI: ； ； The ROI was sized to obtain 128. ROI of 128.

7. The method for identifying a dynamic palmprint ROI based on video according to claim 2, characterized in that, In step S5, palmprint registration is performed based on the palmprint features, including: Obtain the Hamming distance between each keyframe and the ROI of the previous adjacent keyframe, and determine whether the Hamming distance is less than a preset distance threshold. If it is less than the distance threshold, then the current keyframe is used as the registration keyframe. If the total number of processed keyframes reaches a preset total threshold, and the total number of registered keyframes is lower than the number of registered keyframes, then the registration of the palmprint feature fails.

8. A video-based dynamic palmprint ROI recognition system, implementing the steps of the video-based dynamic palmprint ROI recognition method as described in any one of claims 1 to 7, characterized in that, include: The palm video capture module is used to acquire video images of the user's palm and to capture each video frame containing the user's palm. The video frame processing module is used to analyze each video frame one by one and select video frames that contain a hand and have a hand boundary as valid frames. The video frame processing module further analyzes each of the valid frames one by one, obtains the image clarity and image highlight of the valid frames, removes valid frames whose image clarity and image highlight do not meet the preset conditions, and selects the valid frames that meet the preset conditions as key frames. The image preprocessing module is used to obtain palm boundary and finger valley point location information from the preset region in the key frame, extract palm ROI, and extract palm print features; The palmprint recognition module is used to register palmprints based on the palmprint features. The palmprint recognition module acquires real-time video images of the user's palm, acquires the user's palmprint features based on steps S1-S4, and matches the user's palmprint features with the features of the registered palmprints to perform palmprint recognition.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the video-based dynamic palmprint ROI recognition method as described in any one of claims 1 to 7.

Images