Intelligent attendance card punching method and system based on face recognition

Abstract

The invention discloses an intelligent attendance clock-in method and system based on face recognition, and belongs to the technical field of image recognition, and the method comprises the following steps: S1, obtaining each functional region of a clock-in picture; s2, obtaining an optical feedback judgment result of the clock-in picture; s3, if the optical feedback judgment result of a certain clock-in picture is qualified, storing the clock-in picture in an image recognition temporary storage area, and if the optical feedback judgment results of all the clock-in pictures subjected to continuous snapshot are unqualified, carrying out image splicing processing, and storing the processed spliced image in the image recognition temporary storage area; s4, the comprehensive similarity of face recognition is obtained, an attendance checking and card punching judgment result is obtained, prompt information is sent out, face recognition attendance checking can still be rapidly and accurately carried out under the light environment influence condition, and the problem that in the prior art, due to the light environment influence of attendance checking and card punching, the face recognition response speed is low is solved.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to an intelligent attendance tracking method and system based on facial recognition. Background Technology

[0002] Existing intelligent attendance systems use cameras to capture users' facial images in real time, then process these images sequentially through face detection, image preprocessing, and feature extraction, and finally compare them with registered templates in the database to complete identity matching and attendance tracking.

[0003] For example, the intelligent classroom attendance method and system disclosed in Chinese invention patent CN112395950B includes: pre-constructing a face database, including the face image of each student; triggering the electronic class sign and camera in the attendance classroom to enter attendance mode; determining the set of students to be present in the attendance classroom; checking in and out of students through the electronic class sign to obtain first attendance data; periodically collecting student images through the camera to obtain a set of student images; performing face recognition on each image in the set of student images based on the face database to obtain second attendance data; and determining the attendance status of each student in the set of students based on the first attendance data and the second attendance data.

[0004] For example, Chinese invention patent CN113553990B discloses a method, apparatus, computer device, and storage medium for multi-face tracking and recognition, including: acquiring face information of multiple targets within the current monitoring area; for each target, recording the time when the first frame of face image is detected; for each target, determining the acquisition degree of the target's face information; when the acquisition degree does not meet a preset ratio, continuing to track the target's face, supplementing and improving the target's face information to obtain improved target face information; performing face recognition based on the improved target face information to obtain the recognition result; and using the time of the first frame of face image corresponding to the target as the check-in time of the recognition result.

[0005] However, in the process of implementing the inventive technical solution in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems: In existing technologies, attendance tracking is mostly conducted in well-lit indoor environments. However, in outdoor or highly reflective environments, problems such as overexposure or severe shadows in some areas of the face image are often encountered. This can lead to users having to perform face recognition multiple times, and in severe cases, the user may be identified as someone else, seriously affecting the efficiency of face recognition attendance tracking. Therefore, there is a problem of slow face recognition response speed due to the influence of the lighting environment on attendance tracking. Summary of the Invention

[0006] To address the technical problem of slow facial recognition response speed caused by ambient light conditions during attendance tracking, this invention provides an intelligent attendance tracking method and system based on facial recognition. The technical solution is as follows: On the one hand, a smart attendance check-in method based on facial recognition is provided. This method includes: S1. Upon receiving a user's attendance check-in signal, acquiring consecutive snapshots of the user during attendance check-in, statistically analyzing the check-in images, dividing them into regions, and obtaining the functional regions of the check-in images; S2. Acquiring the optical feedback parameters of each dispersed region of the check-in images, analyzing the optical feedback judgment results of the check-in images, where the optical feedback parameters reflect the imaging state of the check-in images; S3. Based on the analysis of the optical feedback judgment results of the check-in images, if the optical feedback judgment result of a certain check-in image is qualified, then the check-in image is stored in the image recognition temporary storage area; if the optical feedback judgment results of all consecutive snapshots of check-in images are unqualified, then image stitching processing is performed, and the processed stitched image is stored in the image recognition temporary storage area; S4. Extracting the stored images from the image recognition temporary storage area, analyzing the comprehensive facial recognition similarity, thereby obtaining the attendance check-in judgment result, and issuing a prompt message.

[0007] On the other hand, a facial recognition-based intelligent attendance system is provided. This system utilizes a facial recognition-based intelligent attendance method and includes: a region division module, an optical feedback analysis module, a stitching processing module, and an attendance determination module. The region division module, upon receiving a user's attendance signal, acquires continuous snapshots of the user's attendance, statistically analyzes the attendance images, and divides them into regions to obtain the functional areas of the attendance images. The optical feedback analysis module acquires the optical feedback parameters of each dispersed region of the attendance images and analyzes them to determine the optical feedback determination of the attendance images. As a result, the optical feedback parameters are used to reflect the imaging status of the attendance images; the stitching processing module is used to analyze the optical feedback judgment results based on the attendance images. If the optical feedback judgment result of a certain attendance image is qualified, the attendance image is stored in the image recognition temporary storage area. If the optical feedback judgment results of all consecutive snapshots of attendance images are unqualified, image stitching processing is performed, and the processed stitched image is stored in the image recognition temporary storage area; the attendance attendance judgment module is used to extract the stored images in the image recognition temporary storage area, analyze them to obtain the comprehensive facial recognition similarity, thereby obtaining the attendance attendance judgment result and issuing a prompt message.

[0008] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: The intelligent attendance method based on face recognition provided by this invention divides attendance images from different angles into regions, analyzes optical feedback parameters, performs image stitching, and calculates comprehensive similarity. This enables the automatic selection or optimization of high-quality face images in complex lighting environments, thereby achieving fast and accurate face recognition attendance even under lighting conditions. This effectively solves the problem of slow face recognition response speed caused by the lighting environment in existing technologies.

[0009] 2. This invention constructs a three-dimensional recognition model and uses facial key point detection combined with PnP pose estimation algorithm to align the three-dimensional model with the check-in image, thereby accurately dividing the area of ​​the check-in image. It also combines three-dimensional point cloud data to achieve one-to-one correspondence of multi-angle image areas, thus realizing accurate matching of facial regions under different angles and complex lighting conditions, and improving the accuracy of facial feature extraction.

[0010] 3. This invention obtains the optical feedback parameters of each dispersed area of ​​the check-in image and makes a judgment based on the optical feedback threshold, thereby automatically distinguishing between qualified and unqualified images. This enables effective evaluation of illumination uniformity and exposure, ensuring the image quality stored in the image recognition temporary storage area and improving the stability and reliability of subsequent face recognition.

[0011] 4. This invention stitches and optimizes dispersed regions, and adjusts the dispersed regions to be optimized by using the average brightness and average exposure of adjacent preferred dispersed regions. This improves image quality even under poor local lighting conditions, thereby achieving adaptive optimization of the comprehensive face image. This ensures that the final stitched second or third check-in image has high recognition usability and effectively improves the accuracy of the comprehensive similarity calculation for face recognition. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart illustrating the intelligent attendance and check-in method based on face recognition provided in this application embodiment; Figure 2 A flowchart outlining the macroscopic steps of the intelligent attendance and check-in method based on face recognition provided in this application embodiment; Figure 3 A flowchart illustrating the image stitching process of the intelligent attendance tracking method based on face recognition provided in this application embodiment; Figure 4 A schematic diagram of the structure of an intelligent attendance system based on face recognition provided in an embodiment of this application; Figure 5 A schematic diagram of adjacent areas for the intelligent attendance check-in method based on face recognition provided in the embodiments of this application; Figure 6 A schematic diagram illustrating the attendance record query of the intelligent attendance system based on face recognition provided in this application embodiment; Figure 7 A schematic diagram of the authentication image quality for the intelligent attendance system based on face recognition provided in this application embodiment; Figure 8 This is a schematic diagram of the attendance verification result of the intelligent attendance system based on face recognition provided in the embodiments of this application. Detailed Implementation

[0014] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0015] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0016] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0017] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] like Figure 1The diagram shows a flowchart of an intelligent attendance check-in method based on face recognition provided in this application embodiment. The method includes the following steps: S1. After receiving a user's attendance check-in signal, acquire consecutive snapshots of the user checking in, statistically obtain check-in images, analyze the check-in images and divide them into regions to obtain functional regions of the check-in images; S2. Acquire optical feedback parameters of each dispersed region of the check-in images, analyze and obtain the optical feedback judgment result of the check-in images, the optical feedback parameters are used to reflect the imaging state of the check-in images; S3. Based on the analysis of the optical feedback judgment result of the check-in images, if the optical feedback judgment result of a certain check-in image is qualified, then the check-in image is stored in the image recognition temporary storage area; if the optical feedback judgment result of all consecutive snapshots of check-in images is unqualified, then image stitching processing is performed, and the processed stitched image is stored in the image recognition temporary storage area; S4. Extract the stored image in the image recognition temporary storage area, analyze and obtain the comprehensive face recognition similarity, thereby obtaining the attendance check-in judgment result, and issuing a prompt message.

[0020] In this embodiment, as Figure 2 As shown, Figure 2 This is a flowchart illustrating the macroscopic steps of the intelligent attendance tracking method based on face recognition provided in this application embodiment. As shown in the diagram, the overall process begins with acquiring attendance images from multiple angles. Then, optical feedback parameters are extracted and analyzed from these images to obtain corresponding optical feedback judgment results. If the judgment result is satisfactory, the image is directly marked as the first attendance image; if the judgment result is unsatisfactory, the process proceeds to the subsequent image stitching step. Both the first attendance image and the stitched attendance image require facial feature extraction, and the comprehensive facial recognition similarity is calculated based on different recognition schemes. Finally, the attendance tracking result is obtained and output, thus completing the entire attendance recognition process.

[0021] This solution acquires check-in images from different angles, specifically images taken during the attendance check-in process. The attendance system (such as an attendance check-in mini-program or attendance machine) will issue voice prompts to remind the user to perform a specified action (such as turning their head left or right). This generates multiple check-in images from the user's perspective. It's important to note that this process captures continuous snapshots of the user during attendance check-in, therefore, it doesn't need to consider any inappropriate user behavior (such as rapidly changing facial expressions or making strange faces). This solution primarily considers individuals performing outdoor work, part-time work, or work in environments with significant changes in lighting conditions (high exposure), such as summer camp part-time workers.

[0022] Based on the attendance check-in result, a prompt message is issued. If the attendance check-in result is successful, a check-in success reminder is issued. If the attendance check-in result is unsuccessful, a check-in failure reminder is issued, and the user is reminded to check in again.

[0023] The attendance check-in result is obtained by: obtaining the preset facial recognition comprehensive similarity threshold in the database and comparing it with the facial recognition comprehensive similarity. If the facial recognition comprehensive similarity is above the facial recognition comprehensive similarity threshold, the attendance check-in result is successful; otherwise, the attendance check-in result is failed.

[0024] Furthermore, the functional regions of the check-in image are obtained. Specifically, the check-in image is input into a 3D recognition model. The 3D recognition model performs facial key point detection on the check-in image and aligns the 3D recognition model with the check-in image based on the PnP pose estimation method. This analysis yields the scattered regions of the check-in image. Each scattered region of the check-in image is then numbered to obtain its unique identifier. Check-in images with the same identifier belong to the same functional region. Thus, the functional regions of the check-in image are statistically determined.

[0025] In this embodiment, it should be noted that the 3D recognition model is a 3D geometric and texture template under a unified face coordinate system. It is typically a dense 3D mesh (vertices / triangles) and semantically divided into several facial functional regions (e.g., left eye region, right eye region, bridge of the nose region, left cheek region, right cheek region, perioral region, jaw region, etc.). This model is used to uniformly map 2D check-in images from different perspectives to the same semantic coordinate system, ensuring the consistency and comparability of "regions with the same name" in different images. Its output format is a parameterized 3D mesh (vertices coordinate array, triangle indices) and a region index for each vertex. The face coordinate system is a local 3D coordinate system (origin, fixed axis) bound to the 3D recognition model, used to define region positions and UV mapping (UV represents 2D texture coordinates) in 3D space. Facial functional regions are predefined semantically dispersed regions on the 3D model or its UV unfolded plane (e.g., the "left eye region" is composed of several vertices / facets on the model). Facial landmark detection specifically involves detecting a set of predefined facial landmarks (such as the corners of the eyes, the tip of the nose, and the corners of the mouth) on a 2D image. This can be achieved using a heatmap regression method based on a convolutional neural network (CNN) (FAN — Face Alignment Network). 3D point cloud data can be obtained through 3D scanning using attendance devices (devices or machines used by users to access attendance tracking apps).

[0026] The specific method for point cloud matching is as follows: acquire the three-dimensional point cloud data of the target face, preprocess the point cloud data using the PCL library, including downsampling using voxel grid filter and removing noise points through statistical outlier removal to improve data quality, extract key points of the point cloud using the ISS (Intrinsic Shape Signatures) algorithm in PCL, and calculate the FPFH (FastPoint Feature Histograms) descriptors of these key points. Using the Open3D library, the RANSAC (Random Sample Consensus) algorithm is used to match feature points of two point clouds to obtain an initial rigid transformation matrix, completing coarse registration. The PCL ICP (Iterative Closest Point) algorithm is then used to perform fine registration on the coarsely registered point clouds. By iteratively optimizing and minimizing the distance between point clouds, the alignment effect of the point clouds is further optimized, ultimately achieving high-precision point cloud matching. Through point cloud matching, regions of the same point cloud data are assigned the same number, for example, the left eye region is numbered 01, the right eye region is numbered 02, the bridge of the nose region is numbered 03, and the perioral region is numbered 04, etc., thus enabling the same regions in check-in images (continuous snapshot images) to be marked with the same number.

[0027] Furthermore, the optical feedback judgment result of the check-in image is obtained. Specifically, the optical feedback parameters of each dispersed region of the check-in image are acquired, including the average brightness, shadow coverage, overexposure area ratio, and brightness range. Based on the optical feedback parameters of each dispersed region of the check-in image and a preset optical feedback calibration set in the database, the optical feedback values ​​of each dispersed region of the check-in image are obtained. The optical feedback calibration set includes the average brightness calibration value, shadow coverage calibration value, overexposure area ratio calibration value, and brightness range calibration value. A preset optical feedback threshold in the database is acquired and compared with the optical feedback values ​​of each dispersed region of the check-in image to obtain the optical feedback judgment result. If the optical feedback values ​​of all dispersed regions of a certain check-in image are above the optical feedback threshold, the optical feedback judgment result of that check-in image is qualified, and that check-in image is marked as the first check-in image; otherwise, the optical feedback judgment result of that check-in image is unqualified.

[0028] In this embodiment, the optical feedback parameters, including the average brightness, shadow coverage, overexposure area ratio, shadow area ratio, and brightness range, can be obtained through the OpenCV image processing tool library (Open Source Computer Vision Library).

[0029] The optical feedback value is obtained by analyzing the optical feedback parameters, taking into account the interrelationships between these parameters. For example, a moderate average brightness provides a clear baseline for face recognition; if the average brightness is too high, the proportion of overexposed areas increases, leading to the loss of facial feature details; if the average brightness is too low, the shadow coverage and shadow area proportion increase, causing blurring of local features. Meanwhile, the brightness range reflects the uniformity of image brightness distribution. A large range often indicates both severe shadows and overexposure, resulting in unstable overall image features. Therefore, the optical feedback value is the result of the interaction of these parameters; a higher optical feedback value indicates better image quality, which is more conducive to the stable extraction and recognition of facial features.

[0030] By acquiring optical feedback parameters (mean brightness, shadow coverage, overexposure area ratio, and brightness range) of various dispersed regions in the check-in image and calculating the optical feedback value for each region, the quality of local facial images can be quantitatively evaluated. This ensures fine-grained analysis of facial region feature differences under different lighting conditions, thereby improving the stability of facial feature extraction. By comparing the optical feedback value with a preset optical feedback calibration set and optical feedback threshold, an optical feedback judgment result is obtained, enabling the classification of check-in images as qualified or unqualified. This effectively ensures the image quality entering the subsequent facial recognition process, improving overall recognition accuracy and response speed.

[0031] Furthermore, the optical feedback values ​​of each dispersed region of the check-in image are obtained. Specifically, the difference between the mean brightness and the calibrated mean brightness is analyzed to obtain the difference determination result; the shadow coverage, reflective area ratio, shadow area ratio, and brightness range are proportionally analyzed with the calibrated shadow coverage, reflective area ratio, shadow area ratio, and brightness range values ​​to obtain the proportional analysis result; based on the difference determination result and the proportional analysis result, corresponding weighting factors are introduced for coupling processing to obtain the optical feedback values ​​of each dispersed region of the check-in image. The optical feedback values ​​are used to represent the usability of the check-in image when extracting facial features.

[0032] In this embodiment, the optical feedback values ​​of each dispersed region of the check-in image are obtained using the following method: In the formula, G i LY represents the optical feedback value of the i-th dispersed region in the check-in image. i Y represents the mean brightness of the i-th dispersed region, LB represents the luminance mean calibration value, and YY i YB represents the shadow coverage of the i-th dispersed region, YS represents the shadow coverage calibration value. iYH represents the overexposure area ratio of the i-th dispersed region, SB represents the overexposure area ratio calibration value, and YH represents the overexposure area ratio. i denoted as the brightness range of the i-th dispersed region, HB represents the brightness range calibration value, ω1 represents the brightness mean weighting factor, ω2 represents the shadow coverage weighting factor, ω3 represents the overexposure area ratio weighting factor, and ω4 represents the brightness range weighting factor.

[0033] It should be noted that the weighting factors for mean brightness, shadow coverage, overexposure area ratio, and brightness range can be obtained from a database. For example, the mean brightness weighting factor can be obtained by retrieving historical mean brightness values ​​stored in the database, comparing them with the mean brightness values, and marking historical mean brightness values ​​whose differences from the mean brightness value are within a preset range as historical control mean brightness values. The weighting factors for each historical control mean brightness value are then obtained and averaged to obtain the mean weighting factor for each historical control mean brightness value. Simultaneously, standard deviation analysis is performed on the weighting factors for each historical control mean brightness value to obtain the standard deviation of the weighting factor for each historical control mean brightness value. A preset standard deviation benchmark value is obtained from the database and compared with the weighting factor for each historical control mean brightness value. The standard deviations of the weighting factors are compared. If the standard deviation of the weighting factor for each historical control brightness mean is below the standard deviation mean, then the average weighting factor for each historical control brightness mean is used as the brightness mean weighting factor. Otherwise, the standard deviation of the weighting factor for each historical control brightness mean is subtracted from the standard deviation benchmark, and the difference is divided by the standard deviation benchmark to obtain the standard deviation adjustment coefficient. The standard deviation adjustment coefficient is multiplied by the average weighting factor for each historical control brightness mean to obtain the standard deviation adjustment value of the weighting factor for each historical control brightness mean. The adjusted weighting factor average is obtained by subtracting the standard deviation adjustment value of the weighting factor for each historical control brightness mean from the average weighting factor average. This adjusted weighting factor average is then used as the weighting factor for the third region identification similarity. Other weighting factors, such as shadow coverage weighting factor, overexposure area ratio weighting factor, and brightness range weighting factor, are obtained in the same way as the brightness mean weighting factor.

[0034] Further, image stitching processing is performed. Specifically, the scattered regions of the check-in image are summarized to obtain the scattered regions of each functional area of ​​the check-in image. The optical feedback value of each scattered region of each functional area of ​​the check-in image is compared with an optical feedback threshold to obtain the stitching selection result for each scattered region. If the optical feedback value of a certain scattered region is above the optical feedback threshold, the stitching selection result for that scattered region is qualified, and that scattered region is marked as a qualified scattered region. If the optical feedback value of a certain scattered region is below the optical feedback threshold, the stitching of that scattered region is not allowed. If the optional judgment result is unqualified, the dispersed area is marked as a candidate dispersed area. Based on the analysis of each dispersed area, the splicing judgment result of each functional area attribute is obtained. Based on the splicing judgment result of each functional area attribute, the corresponding dispersed area optimization processing is performed to obtain each preferred dispersed area of ​​the check-in image. Then, image splicing processing is performed to obtain the spliced ​​check-in image. If the spliced ​​check-in image is composed entirely of optional qualified dispersed areas, then the check-in image is the second check-in image. If there are candidate dispersed areas in the spliced ​​check-in image, then the check-in image is the third check-in image.

[0035] In this embodiment, Figure 3 The image stitching process flowchart for the intelligent attendance check-in method based on face recognition provided in this application embodiment is as follows: First, the optical feedback values ​​of each dispersed region in the check-in image are obtained and compared with a preset optical feedback threshold to obtain a stitching selection result. If a dispersed region is deemed qualified, it is marked as a qualified dispersed region; if it is deemed unqualified, it is marked as a candidate dispersed region. Next, based on the dispersed region results, functional region attribute stitching is determined. If there are qualified dispersed regions in the functional region, the dispersed region with the largest feedback value is directly selected as the preferred region; if all functional regions are unqualified, the dispersed region with the largest feedback value is selected as the region to be optimized. The region to be optimized needs to further obtain the average brightness and exposure of adjacent preferred regions, and adjust its brightness and exposure to obtain the optimized preferred dispersed region. All preferred dispersed regions are stitched together to obtain the stitched image. If the stitched image consists entirely of qualified dispersed regions, the stitched image is the second check-in image; if the stitched image contains optimized regions, the stitched image is the third check-in image.

[0036] It's important to note that stitching images of different functional regions to obtain a complete face recognition image is advantageous compared to not stitching when extracting and analyzing features from different images to calculate overall similarity. Without stitching, different functional regions may be scattered across multiple images; for example, the forehead region in the upper left corner and the mouth region in the lower right corner may exist in different images (referring to clear images of these regions). This makes it impossible to calculate the Euclidean distance between these distant feature points on the same image, resulting in limited local similarity and restricting the overall similarity assessment. By stitching, the scattered regions are integrated into the same image according to their functional regions, mapping all key regions to the same coordinate system. This allows for the calculation of feature Euclidean distances between any regions on the same image, obtaining more and more complete feature similarity information. This makes the calculation of overall similarity more comprehensive and the feature coverage more sufficient, more accurately reflecting the overall facial similarity between different images, thus significantly improving the accuracy and stability of face recognition. Furthermore, the stitched image can effectively avoid recognition errors caused by uneven lighting, partial occlusion, or low-quality areas in a single image, improving the stability of the overall similarity of face recognition, thereby significantly improving recognition performance and attendance accuracy under different lighting and pose conditions.

[0037] When stitching together multiple check-in photos (including consecutively taken images), due to different shooting angles, some images may show side profiles while others show frontal faces. This can cause the same functional areas to appear stretched or scaled in different images, affecting the accuracy of subsequent face recognition. Therefore, stitching correction can be performed based on 3D point cloud data. The specific process is as follows: First, obtain the 3D point cloud data of the face corresponding to each image through the check-in device, and align each point cloud to a unified face coordinate system using a 3D recognition model. Second, perform normal vector and depth analysis on the point cloud corresponding to the functional area of ​​each image to determine the degree of deflection or tilt caused by different angles. Then, standardize the point cloud of each functional area through rigid transformations (rotation, translation, scaling) to ensure that functional areas at different angles maintain consistent size and proportion under a unified coordinate system. Finally, stitch the corrected 2D images according to the point cloud mapping positions to ensure that each functional area maintains the correct proportion and position in the stitched image, thereby avoiding local magnification or shrinkage caused by angle differences and improving the accuracy of face recognition in the stitched image.

[0038] By comparing the scattered regions of the check-in image with optical feedback thresholds and dividing the results into selectable and alternative scattered regions, quality screening of local areas can be achieved. This allows for the priority use of qualified regions during image stitching, ensuring that the overall quality of the stitched image is significantly better than the original image. Based on the stitching judgment results of each functional region, differentiated optimization processing is applied to scattered regions of different qualities. Preferred scattered regions are then stitched together to avoid the loss of certain facial features due to poor quality in a single image. This achieves comprehensive preservation of facial features and ensures the accuracy of subsequent feature extraction.

[0039] This invention, by dynamically determining and optimizing the differences in optical feedback values ​​during the stitching process, effectively mitigates the impact of uneven lighting, local overexposure, or shadow coverage on the overall image recognition performance, achieving stability of the stitched image under different lighting environments and thus improving recognition robustness. By distinguishing between the second check-in image (completely stitched from qualified dispersed regions) and the third check-in image (including stitched from alternative dispersed regions), this invention provides a clear quality level identifier for subsequent recognition stages. This allows the recognition module to adjust feature extraction and comparison strategies based on the stitched image level, improving the recognition response speed and success rate of face recognition under different quality conditions. Existing technologies often deem images invalid due to substandard overall image quality, wasting information from areas with good quality and affecting face recognition performance. This invention, through region-level stitching and optimization, enables the generation of usable face images even in areas with low quality within the overall image, effectively improving image utilization and avoiding duplicate check-ins and delays caused by invalid data collection.

[0040] Furthermore, the splicing judgment results of each functional area are obtained. The specific method is as follows: based on the analysis of each dispersed area of ​​each functional area, if the optical feedback value of a dispersed area of ​​a certain functional area is a selectable qualified dispersed area, then the splicing judgment result of the functional area is qualified; if each dispersed area of ​​a certain functional area is a candidate dispersed area, then the splicing judgment result of the functional area is unqualified.

[0041] In this embodiment, the present invention comprehensively analyzes the optical feedback values ​​of each dispersed region at the functional region level, avoiding misjudgment of the overall functional region due to anomalies in individual dispersed regions. This achieves a more robust functional region quality assessment and improves the accuracy of stitching judgment. The present invention allows for some dispersed regions to be unqualified in functional region judgment. As long as qualified dispersed regions still exist within the functional region, the functional region can be judged as qualified, and qualified images are selected for stitching. This effectively avoids the impact of small-scale optical anomalies on the overall image stitching effect.

[0042] Furthermore, the preferred dispersed regions of the check-in image are obtained through the following method: based on the splicing judgment results of each functional region attribute, if the splicing judgment result of a certain functional region attribute is qualified, then the dispersed region corresponding to the maximum optical feedback value of that functional region is selected as the preferred dispersed region of that functional region; if the splicing judgment result of a certain functional region attribute is unqualified, then the dispersed region corresponding to the maximum optical feedback value of that functional region is selected, marked as the dispersed region to be optimized, and the dispersed region to be optimized is subjected to dispersion region optimization processing to obtain the optimized dispersed region, which is then marked as the preferred dispersed region. Thus, the preferred dispersed regions of each functional region are obtained, and then spliced ​​together to obtain the processed spliced ​​image, which is then marked as the third check-in image.

[0043] In this embodiment, by selecting the dispersion region corresponding to the maximum optical feedback value as the preferred dispersion region when the functional region is qualified, the optimal region selection for that functional region is achieved, ensuring that the image quality of the stitched image in key areas is always at its best. When the functional region is unqualified, the region is not directly discarded, but the dispersion region with the maximum feedback value is optimized and transformed into a preferred dispersion region, thereby ensuring the integrity of the functional region during stitching and avoiding the problem of insufficient facial feature information caused by local missing parts. By directly selecting the best from qualified regions and optimizing the best from unqualified regions, the overall consistency of the stitched image can be guaranteed, effectively reducing visual disharmony caused by abnormalities such as excessive brightness or darkness in certain areas, thereby improving the accuracy and stability of subsequent feature extraction. Thus, under complex conditions such as different lighting, different angles, and even makeup occlusion, the dispersion region can be dynamically selected or optimized based on the optical feedback value, thereby forming a reliable stitched image, which significantly enhances the adaptability and robustness of the system under complex working conditions. Since the stitched image is always composed of optimized dispersed regions, the overall image quality is guaranteed to be optimal, thereby improving the accuracy of facial feature extraction. This effectively reduces the false recognition rate caused by poor image quality, while also speeding up the recognition response and improving attendance efficiency.

[0044] Furthermore, the dispersed area optimization process is performed. Specifically, the preferred dispersed areas adjacent to the dispersed area to be optimized are obtained and marked as adjacent preferred dispersed areas. The brightness and exposure of each adjacent preferred dispersed area are obtained and averaged to obtain the average brightness and exposure of the adjacent preferred dispersed areas. Based on the average brightness and exposure of the adjacent preferred dispersed areas, the brightness and exposure of the dispersed area to be optimized are adjusted.

[0045] In this embodiment, as Figure 5 As shown, Figure 5This diagram illustrates the adjacent regions of the intelligent attendance tracking method based on face recognition provided in this application embodiment. As shown, if region A is the dispersed region to be optimized, then regions B, C, D, and E adjacent to region A are analyzed. If there are preferred dispersed regions in regions B, C, D, and E, then the preferred regions in regions B, C, D, and E are the adjacent preferred dispersed regions, thus obtaining each adjacent preferred dispersed region. If there are no preferred regions in regions B, C, D, and E, then region A and regions B, C, D, and E are considered as a whole region, labeled as region combination ABCDE, and its adjacent regions are analyzed again. At this time, the adjacent regions of region combination ABCDE are regions F, G, H, I, J, K, L, and M, and the process continues to determine whether there are preferred regions in regions F, G, H, I, J, K, L, and M until the preferred dispersed regions are found.

[0046] The brightness and exposure of the area to be optimized are adjusted based on the average brightness and average exposure of adjacent preferred dispersion areas. Specifically, the brightness value of the optimized dispersion area is adjusted to the average brightness of adjacent preferred dispersion areas, and the exposure of the optimized dispersion area is adjusted to the average exposure of adjacent preferred dispersion areas.

[0047] By acquiring the brightness and exposure values ​​of each preferred dispersed region adjacent to the dispersed region to be optimized, and averaging them, the dispersed region to be optimized is then adjusted accordingly. This achieves adaptive brightness and exposure optimization of the dispersed region, ensuring reasonable processing of the visual effect of the local area. As a result, the stitched image appears more natural in terms of facial structure information, which is beneficial for subsequent facial key point detection and feature extraction.

[0048] Furthermore, the comprehensive facial recognition similarity is obtained. Specifically, the optimized check-in images from the image recognition temporary storage area are extracted and analyzed using a preset facial recognition scheme to obtain the comprehensive facial recognition similarity. The optimized check-in images include a first check-in image, a second check-in image, and a third check-in image. The preset facial recognition scheme includes a first facial recognition scheme, a second facial recognition scheme, and a third facial recognition scheme. If the optimized check-in image is the first check-in image, then the first check-in image is analyzed for facial recognition features using the first facial recognition scheme to obtain the comprehensive facial recognition similarity. The first facial recognition scheme specifically involves: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] The system extracts overall image features from the check-in image to obtain the first face recognition feature. It then retrieves a pre-stored face recognition feature set from the database and performs feature similarity analysis with the first face recognition feature to obtain the similarity between the first check-in image and the face recognition feature set, marking it as the comprehensive face recognition similarity. If the check-in image is optimized to become a second check-in image, the second check-in image is then subjected to face recognition feature analysis using a second face recognition scheme to obtain the comprehensive face recognition similarity. The second face recognition scheme specifically involves extracting features from each preferred dispersed region of the second check-in image to obtain the second face recognition feature of each preferred dispersed region. Face recognition features are analyzed and compared with a face recognition feature set to obtain the face recognition similarity between each preferred dispersed region and the face recognition feature set. The face recognition similarity between each preferred dispersed region and the face recognition feature set is then averaged to obtain the comprehensive face recognition similarity. If the optimized check-in image is the third check-in image, then the third check-in image is analyzed using a third face recognition scheme to obtain the comprehensive face recognition similarity. The third face recognition scheme specifically involves: extracting features from each preferred dispersed region of the third check-in image to obtain the third face recognition features of each preferred dispersed region. The system extracts features from the third check-in image and performs feature similarity analysis with the face recognition feature set to obtain the face recognition similarity between each preferred dispersed region and the face recognition feature set. The face recognition similarity between each preferred dispersed region and the face recognition feature set is then averaged to obtain the third region recognition similarity. The system extracts overall image features from the third check-in image and performs feature similarity analysis with the face recognition feature set to obtain the similarity between the third check-in image and the face recognition feature set, and marks it as the third overall recognition similarity. The third region recognition similarity and the third overall recognition similarity are coupled and analyzed to obtain the comprehensive face recognition similarity.

[0049] In this embodiment, a hierarchical face recognition scheme employs overall feature extraction, dispersed region feature extraction, and region-to-whole coupling analysis for optimized check-in images of varying quality. This approach adaptively selects the matching recognition method based on the quality differences between the first, second, and third check-in images, achieving stable and accurate overall face recognition similarity even with varying image quality. This method ensures rapid recognition under high-quality images while enhancing robustness in low-quality images or when some regions require optimization through the use of mean region features and region-to-whole coupling. This effectively improves the accuracy and reliability of face recognition under complex lighting and diverse scenarios.

[0050] A second face recognition scheme is used to analyze face recognition features and obtain the overall face recognition similarity. The specific method is as follows: ; In the formula, XS represents the overall similarity score of face recognition, and S... j This represents the face recognition similarity of the j-th preferred dispersed region, where j represents the number of the preferred dispersed region, j=1,2,...,j max ,j max This represents the total number of preferred dispersed areas.

[0051] A third-party facial recognition scheme is used to analyze facial features and obtain the overall facial similarity score. The specific method is as follows: ; In the formula, XS represents the overall similarity score of face recognition, and S... j This represents the face recognition similarity of the j-th preferred dispersed region, where j represents the number of the preferred dispersed region, j=1,2,...,j max ,j max ε1 represents the total number of preferred dispersed regions, SA represents the third overall identification similarity, ε1 represents the weighting factor of the third region identification similarity, and ε2 represents the weighting factor of the third overall identification similarity.

[0052] The weighting factors for the third region identification similarity and the third overall identification similarity can be obtained from a database. For example, historical third overall identification similarities stored in the database can be retrieved and compared with the third overall identification similarity. Historical third overall identification similarities whose differences from the third overall identification similarity are within a preset range are marked as historical control third overall identification similarities. The weighting factors of each historical control third overall identification similarity are obtained and mean-valued to obtain the mean of the weighting factors of each historical control third overall identification similarity. At the same time, standard deviation analysis is performed on the weighting factors of each historical control third overall identification similarity to obtain the standard deviation of the weighting factors of each historical control third overall identification similarity. The preset standard deviation benchmark value in the database is obtained and compared with the standard deviation of the weighting factors of each historical control third overall identification similarity. For comparison, if the standard deviation of the weighting factor for the third overall identification similarity of each historical control is below the mean standard deviation, then the mean weighting factor for the third overall identification similarity of each historical control is used as the weighting factor for the third region identification similarity. Otherwise, the standard deviation of the weighting factor for the third overall identification similarity of each historical control is compared with the standard deviation benchmark value, and the difference is divided by the standard deviation benchmark value to obtain the standard deviation adjustment coefficient. The standard deviation adjustment value of the weighting factor for the third overall identification similarity of each historical control is obtained by multiplying the standard deviation adjustment coefficient by the mean weighting factor for the third overall identification similarity of each historical control. The adjusted weighting factor mean is obtained by subtracting the adjusted weighting factor from the mean weighting factor for the third overall identification similarity of each historical control, and this adjusted weighting factor mean is used as the weighting factor for the third region identification similarity.

[0053] like Figure 6 As shown, Figure 6 This diagram illustrates the query of attendance records for an intelligent attendance system based on facial recognition, as provided in this application embodiment. The diagram shows the attendance time, location, duration, and status of the person clocking in. It also shows the attendance photo of the person clocking in and the attendance photo in the system template.

[0054] like Figure 7 , Figure 8 As shown, Figure 7 This is a schematic diagram illustrating the image quality for identity verification in the intelligent attendance system based on facial recognition provided in this embodiment of the application. Figure 8This diagram illustrates the identity verification results of a smart attendance system based on facial recognition, as provided in this application embodiment. The diagram shows the image quality of each functional area of ​​the attendance clerk, the similarity between the attendance clerk's image and the attendance clerk's image in the system template, the recognition time of the attendance operation, the total attendance time, and the facial recognition verification results, location verification results, network status, and server response status during attendance.

[0055] like Figure 4 As shown, Figure 4 This is a schematic diagram of the structure of a face recognition-based intelligent attendance system provided in an embodiment of this application. The system includes: a region division module, an optical feedback analysis module, a stitching processing module, and an attendance determination module. The region division module, upon receiving a user's attendance signal, acquires continuous snapshots of the user's attendance, statistically analyzes the attendance images, and divides them into regions to obtain functional areas of the attendance images. The optical feedback analysis module acquires the optical feedback parameters of each dispersed region of the attendance images and analyzes the optical properties of the attendance images. The system provides feedback on the judgment results. Optical feedback parameters reflect the imaging status of the attendance image. The stitching processing module analyzes the optical feedback judgment results based on the attendance images. If the optical feedback judgment result of a particular attendance image is qualified, the image is stored in the image recognition temporary storage area. If the optical feedback judgment results of all consecutive snapshots of attendance images are unqualified, image stitching is performed, and the processed stitched image is stored in the image recognition temporary storage area. The attendance judgment module extracts the stored images from the image recognition temporary storage area and analyzes them to obtain the comprehensive facial recognition similarity, thereby obtaining the attendance judgment result and issuing a prompt message.

[0056] In summary, this embodiment, by dividing check-in images from different angles into regions, analyzing optical feedback parameters, performing image stitching, and calculating comprehensive similarity, can automatically filter or optimize high-quality facial images in complex lighting environments. This enables fast and accurate facial recognition attendance even under lighting conditions, effectively solving the problem of slow facial recognition response speed caused by lighting environment influences in existing technologies.

[0057] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0058] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0060] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0061] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0062] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A smart attendance tracking method based on facial recognition, characterized in that, Includes the following steps: S1. After receiving the user's attendance check-in signal, obtain continuous snapshots of the user's attendance check-in, collect the check-in images, analyze the check-in images and divide them into regions to obtain the functional regions of the check-in images. S2. Obtain the optical feedback parameters of each dispersed area of ​​the check-in image, and analyze the optical feedback judgment result of the check-in image. The optical feedback parameters are used to reflect the imaging state of the check-in image. S3. Based on the optical feedback judgment result analysis of the check-in image, if the optical feedback judgment result of a certain check-in image is qualified, the check-in image is stored in the image recognition temporary storage area. If the optical feedback judgment result of all consecutive check-in images is unqualified, image stitching is performed, and the processed stitched image is stored in the image recognition temporary storage area. S4. Extract the stored image from the image recognition temporary storage area and analyze it to obtain the comprehensive similarity of the face recognition, thereby obtaining the attendance check-in judgment result and issuing a prompt message.

2. The intelligent attendance and check-in method based on face recognition as described in claim 1, characterized in that: The specific method for obtaining the functional areas of the check-in image is as follows: The check-in image is input into a 3D recognition model. The 3D recognition model performs facial landmark detection on the check-in image and aligns the 3D recognition model with the check-in image based on the PnP pose estimation method. This analysis yields the scattered regions of the check-in image, and each scattered region is numbered. Check-in images with the same number belong to the same functional region, thus statistically determining the functional regions of the check-in image.

3. The intelligent attendance and check-in method based on face recognition as described in claim 1, characterized in that: The specific method for obtaining the optical feedback determination result of the check-in image is as follows: Obtain optical feedback parameters for each dispersed region of the check-in image, including average brightness, shadow coverage, overexposure area ratio, and brightness range. The optical feedback values ​​of each dispersed region of the check-in image are obtained by analyzing the optical feedback parameters of each dispersed region of the check-in image with the preset optical feedback calibration set in the database. The optical feedback calibration set includes average brightness calibration value, shadow coverage calibration value, overexposure area ratio calibration value, and brightness range calibration value; Obtain the preset optical feedback threshold from the database and compare it with the optical feedback values ​​of each dispersed area of ​​the check-in image to obtain the optical feedback judgment result of the check-in image. If the optical feedback values ​​of each dispersed area of ​​a certain check-in image are all above the optical feedback threshold, the optical feedback judgment result of the check-in image is qualified and the check-in image is marked as the first check-in image; otherwise, the optical feedback judgment result of the check-in image is unqualified.

4. The intelligent attendance and check-in method based on face recognition as described in claim 3, characterized in that: The specific method for obtaining the optical feedback values ​​of each dispersed region of the check-in image is as follows: The difference between the mean brightness value and the calibrated mean brightness value is analyzed to obtain the result of the difference determination. The shadow coverage, reflective area ratio, shadow area ratio, and luminance range were compared with the calibration values ​​of shadow coverage, reflective area ratio, shadow area ratio, and luminance range, respectively, to obtain the ratio analysis results. Based on the difference degree determination results and the ratio analysis results, corresponding weighting factors are introduced for coupling processing to obtain the optical feedback values ​​of each dispersed region of the check-in image. The optical feedback values ​​are used to represent the usability of the check-in image when extracting facial features.

5. The intelligent attendance and check-in method based on face recognition as described in claim 1, characterized in that: The specific method for image stitching processing is as follows: The scattered areas of the check-in image are summarized to obtain the scattered areas of each functional area of ​​the check-in image. The optical feedback values ​​of each scattered area in each functional area of ​​the check-in image are compared with the optical feedback threshold to obtain the splicing selection result of each scattered area. If the optical feedback value of a certain scattered area is above the optical feedback threshold, the splicing selection result of the scattered area is qualified and the scattered area is marked as a qualified scattered area. If the optical feedback value of a certain scattered area is less than the optical feedback threshold, the splicing selection result of the scattered area is unqualified and the scattered area is marked as a candidate scattered area. Based on the analysis of each dispersed region, the splicing judgment results of each functional region attribute are obtained. Based on the splicing judgment results of each functional region attribute, the corresponding dispersed region optimization processing is performed to obtain each preferred dispersed region of the check-in image. Then, image splicing processing is performed to obtain the spliced ​​check-in image. If the spliced ​​check-in image is composed entirely of selectable qualified dispersed regions, then the check-in image is the second check-in image. If there are candidate dispersed regions in the spliced ​​check-in image, then the check-in image is the third check-in image.

6. The intelligent attendance and check-in method based on face recognition as described in claim 5, characterized in that: The specific method for obtaining the attribute splicing determination results of each functional area is as follows: Based on the analysis of the scattered areas of each functional area, if the optical feedback value of a scattered area of ​​a certain functional area is a selectable qualified scattered area, then the attribute splicing judgment result of that functional area is qualified; if all scattered areas of a certain functional area are candidate scattered areas, then the attribute splicing judgment result of that functional area is unqualified.

7. The intelligent attendance and check-in method based on face recognition as described in claim 5, characterized in that: The specific method for obtaining the preferred dispersed regions of the check-in image is as follows: Based on the analysis of the splicing judgment results of each functional area attribute, if the splicing judgment result of a certain functional area attribute is qualified, then the dispersed area corresponding to the maximum optical feedback value of that functional area is selected as the preferred dispersed area of ​​that functional area. If the splicing result of a certain functional area is unqualified, the scattered area corresponding to the maximum optical feedback value of that functional area is selected and marked as the scattered area to be optimized. The scattered area to be optimized is then processed to obtain the optimized scattered area, which is then marked as the preferred scattered area. Thus, the preferred scattered areas of each functional area are obtained and spliced ​​together to obtain the processed spliced ​​image, which is then marked as the third check-in image.

8. The intelligent attendance and check-in method based on face recognition as described in claim 7, characterized in that: The specific method for performing the decentralized region optimization processing is as follows: Obtain each preferred dispersion region adjacent to the dispersion region to be optimized, and mark them as each adjacent preferred dispersion region. Obtain the brightness value and exposure of each adjacent preferred dispersion region, and perform mean processing to obtain the mean brightness value and mean exposure value of the adjacent preferred dispersion regions. The brightness and exposure of the area to be optimized are adjusted based on the average brightness and average exposure of adjacent preferred dispersion areas.

9. The intelligent attendance and check-in method based on face recognition as described in claim 1, characterized in that: The method for obtaining the comprehensive facial recognition similarity is as follows: Optimized check-in images are extracted from the image recognition temporary storage area and analyzed through a preset face recognition scheme to obtain the comprehensive face recognition similarity. The optimized check-in images include a first check-in image, a second check-in image, and a third check-in image. The preset face recognition scheme includes a first face recognition scheme, a second face recognition scheme, and a third face recognition scheme; If the optimized check-in image is the first check-in image, then the first check-in image is subjected to facial recognition feature analysis using the first facial recognition scheme to obtain the comprehensive facial recognition similarity. The first face recognition scheme is as follows: extract the overall image features of the first check-in image to obtain the first face recognition features, obtain the face recognition feature set pre-stored in the database, and perform feature similarity analysis with the first face recognition features to obtain the similarity between the first check-in image and the face recognition feature set, and mark it as the face recognition comprehensive similarity. If the optimized check-in image is the second check-in image, then the second check-in image is subjected to facial recognition feature analysis using the second facial recognition scheme to obtain the comprehensive facial recognition similarity. The second face recognition scheme is as follows: each preferred dispersed region of the second check-in image is subjected to feature extraction to obtain the second face recognition feature of each preferred dispersed region, and feature similarity analysis is performed with the face recognition feature set to obtain the face recognition similarity between each preferred dispersed region and the face recognition feature set. The face recognition similarity between each preferred dispersed region and the face recognition feature set is averaged to obtain the comprehensive face recognition similarity. If the optimized check-in image is the third check-in image, then the third check-in image is subjected to facial recognition feature analysis using a third facial recognition scheme to obtain the comprehensive facial recognition similarity. The third face recognition scheme is as follows: extract features from each preferred dispersed region of the third check-in image to obtain the third face recognition features of each preferred dispersed region, and perform feature similarity analysis with the face recognition feature set to obtain the face recognition similarity between each preferred dispersed region and the face recognition feature set. Then, average the face recognition similarity between each preferred dispersed region and the face recognition feature set to obtain the third region recognition similarity. The overall image features of the third check-in image are extracted and compared with the face recognition feature set to obtain the similarity between the third check-in image and the face recognition feature set, and it is marked as the third overall recognition similarity. By coupling the third region recognition similarity and the third overall recognition similarity, the comprehensive face recognition similarity is obtained.

10. A system applying the intelligent attendance and check-in method based on face recognition as described in any one of claims 1-9, characterized in that, include: The module includes a region division module, an optical feedback analysis module, a stitching processing module, and an attendance check-in / check-out judgment module. The region division module is used to obtain continuous snapshots of the user's attendance check-in after receiving the user's attendance check-in signal, to obtain the check-in images, to analyze the check-in images and divide them into regions, and to obtain the functional regions of the check-in images. The optical feedback analysis module is used to acquire the optical feedback parameters of each dispersed area of ​​the check-in image, and analyze them to obtain the optical feedback judgment result of the check-in image. The optical feedback parameters are used to reflect the imaging state of the check-in image. The stitching processing module is used to analyze the optical feedback judgment result based on the check-in image. If the optical feedback judgment result of a certain check-in image is qualified, the check-in image is stored in the image recognition temporary storage area. If the optical feedback judgment result of all consecutive check-in images is unqualified, the image stitching processing is performed, and the processed stitched image is stored in the image recognition temporary storage area. The attendance check-in determination module is used to extract the stored image in the image recognition temporary storage area, analyze it to obtain the comprehensive similarity of the face recognition, thereby obtaining the attendance check-in determination result and issuing a prompt message.

Images