A personnel management and control method, system and device based on face recognition

Abstract

This application relates to a method, system, and device for personnel management based on facial recognition. The method is deployed within a facial recognition device, which is an integrated unit containing an electronic screen and processor for human-computer interaction. The method acquires facial image data captured by a facial capture camera and a video surveillance camera connected to the facial recognition device. It filters the facial image data according to facial image quality rules to obtain target facial image data that meets these rules. The target facial image data is compared with facial features in a database to generate a facial recognition result. Strangers are then documented, and combined with age recognition, the results are applied to scenarios requiring registration for entry and to places where minors are prohibited or restricted from entering. This addresses the problem of how to improve personnel management in related technologies.

Description

Technical Field

[0001] This application relates to the technical field of computer vision, and in particular to a method, system and device for personnel management based on face recognition. Background Technology

[0002] Facial recognition, as an important identity verification technology, has been widely used in security, access control, attendance, and personnel management in specific locations. It is particularly useful in edge scenarios requiring real-time identity verification and dynamic control, such as the front desks of hotels, attendance centers, and entertainment venues.

[0003] Existing edge-based facial recognition-based personnel management methods typically employ an architecture based on small, "box-like" edge servers. This method detects, captures, and recognizes faces by accessing video streams. However, this approach has significant limitations: First, at the hardware deployment and interaction level, this architecture usually relies on clients accessing and operating the system through a web browser on a personal computer, resulting in a cumbersome and inefficient facial recognition process. Simultaneously, the server needs to be placed in a server room or mounted on a wall, requiring additional peripherals such as monitors, keyboards, and mice for system deployment and configuration, leading to low deployment efficiency, high costs, and difficulty in flexible and convenient use in various front-end desktop scenarios. Second, at the data collection level, existing technologies collect data through a single collection device, resulting in inaccuracies in recognition accuracy. Finally, at the business functionality level, existing methods are limited in function and struggle to perform personnel management based on recognition results. For example, they are unable to statistically analyze and issue warnings about the frequency of specific strangers appearing in sensitive areas; they are unable to analyze the length of stay and departure status of key personnel; and they are unable to identify minors entering places where minors are prohibited or restricted.

[0004] Therefore, no effective solution has yet been proposed for improving the level of personnel management based on existing methods. Summary of the Invention

[0005] This application provides a method, system, and device for personnel management based on facial recognition, in order to at least address the problem of how to improve the level of personnel management in related technologies.

[0006] In a first aspect, embodiments of this application provide a personnel management method based on facial recognition. The method is deployed on a facial recognition device, which is an integrated unit comprising an electronic screen and a processor for human-computer interaction. The method includes: The face image data collected by the face capture camera and video surveillance camera connected to the face recognition device is obtained. The face image data is filtered according to face image quality rules to obtain target face image data that conforms to the face image quality rules; The target face image data is compared with the face features in the database to generate a face recognition result; Personnel management is carried out based on the facial recognition results.

[0007] In one embodiment, the step of acquiring facial image data captured by the face capture camera and the video surveillance camera connected through the face recognition device includes: The video surveillance camera is connected to the face recognition device, and video stream data is acquired through the video surveillance camera; according to the face capture algorithm, the face target in the video stream data is continuously tracked, and according to the first preliminary face image quality rule, a video face image conforming to the first preliminary face image quality rule is acquired; based on the video face image, video face attribute information is analyzed, and the video face image and the video face attribute information are used as the face image data acquired by the video surveillance camera; By accessing the face capture camera, the system receives captured face images collected by the face capture camera; according to the second preliminary face image quality rules, it obtains captured face images that conform to the second preliminary face image quality rules; based on the captured face images, it analyzes the captured face attribute information, and uses the captured face images and the captured face attribute information as face image data collected by the face capture camera.

[0008] In one embodiment, the step of continuously tracking the facial target in the video stream data and obtaining a video facial image that conforms to the first preliminary facial image quality rule includes: Continuous detection and tracking of target faces in video stream sequences; During the tracking process, the face image is evaluated in real time based on the first preliminary face image quality rules until a preset trigger event occurs. The quality evaluation is based on attributes including the rate of change of the face detection bounding box size. The trigger event includes at least one of the following: The target face leaves the video frame; The duration of continuous tracking of the target face exceeds a preset tracking timeout threshold; The number of video face images that meet the first preliminary face image quality rules exceeds the preset number; In response to the instruction of receiving the trigger event, the video face image that conforms to the first preliminary face image quality rules and the corresponding face attribute information are used as the video face image.

[0009] In one embodiment, the step of filtering the face image data according to face image quality rules to obtain target face image data that conforms to the face image quality rules includes: The facial image data is comprehensively evaluated based on preset facial image quality rules. The quality rules include at least: three-dimensional face angle, image brightness, image clarity, face integrity, face confidence, and the occlusion status of the eyes, mouth, and nose. The evaluation results of the quality rule items are weighted according to preset weights to obtain the comprehensive quality score of the face image data. The overall quality score is compared with a preset overall quality threshold, and face image data that is greater than or equal to the overall quality threshold is marked as target face image data that conforms to the face image quality rules.

[0010] In one embodiment, the step of performing a comprehensive quality assessment of the face image data based on preset face image quality rules includes: Set an independent judgment threshold for each of the quality rule items; Determine whether the evaluation results of the face image data on all the quality rule items meet the corresponding independent judgment thresholds; If so, the facial image data is determined to have passed the comprehensive quality assessment; If not, the facial image data is determined to have failed the comprehensive quality assessment.

[0011] In one embodiment, comparing the target face image data with face features in a database to generate a face recognition result includes: Facial features are extracted from the target face image data to generate a feature vector to be compared; The similarity between the feature vector to be compared and the N face feature vectors pre-stored in the database is calculated to obtain N similarity scores; Select the database personnel information corresponding to the highest similarity score from N similarity scores; The highest score is compared with a preset threshold, and the face recognition result is determined based on the comparison result and the personnel information in the database.

[0012] In one embodiment, comparing the highest score with a preset threshold and determining the face recognition result based on the comparison result and the personnel information in the database includes: If the highest score is greater than or equal to the preset threshold, the recognition is determined to be successful, and the database personnel information corresponding to the highest score is output as the face recognition result. If the highest score is less than the preset threshold, the recognition is deemed to have failed, and the target face image data is marked as stranger information as the face recognition result.

[0013] In one embodiment, personnel control based on the facial recognition results includes: Information on individuals identified as strangers is automatically archived, and their frequency of appearance in different time periods and regions is statistically analyzed. When the frequency of a stranger's appearance exceeds a preset threshold, an early warning mechanism is triggered. Based on continuous facial recognition records, the system calculates the duration of a specific person's stay in the monitored area. When the stay exceeds a preset time limit, an early warning mechanism is triggered. Based on the identified age attributes, minors are initially screened out, and at least one of the following operations is performed: Based on the manually entered name and ID number, obtain the ID information, and submit the ID information and a captured face photo or a face photo collected in cooperation by calling the trusted identity authentication interface provided by the public security department or a third party, and request the return of the verification result; Based on the connection of the ID card reader, the system automatically reads the ID card information and compares the portrait photo in the ID card information with the on-site captured face photo or the face photo collected in cooperation with the system's local algorithm to obtain the verification result; or it submits the ID card information and the on-site captured face photo or the face photo collected in cooperation with the system by calling the trusted identity authentication interface provided by the public security department or a third party, and requests the return of the verification result. After verification, the minor's identity information, facial photo, and recognition record will be linked and stored to complete the minor's information registration.

[0014] Secondly, embodiments of this application provide a personnel management system based on face recognition. The system is used to execute the method described in the first aspect above. The system includes a face image data module, a target face image data module, a face recognition module, and a personnel management module; wherein: The face image data module is used to acquire face image data collected by the face capture camera and the video surveillance camera connected through the face recognition device. The target face image data module is used to filter the face image data according to face image quality rules to obtain target face image data that conforms to the face image quality rules. The face recognition module is used to compare the target face image data with face features in the database to generate a face recognition result; The personnel management module is used to manage personnel based on the facial recognition results.

[0015] Thirdly, embodiments of this application provide a personnel management device based on facial recognition, comprising: Electronic screens are used for human-computer interaction; Communication module, user network communication; Cameras are used to collect facial information of targets for verification and / or registration; The USB interface connects to an external ID card reader, which is used to collect and verify the ID card information of the personnel. Combined with the security module of the facial recognition software, it can achieve secure protection for data transmission and storage. A memory and a processor, a program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements a face recognition-based personnel management method as described in any one of claims 1 to 8.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a personnel management method based on face recognition as described in the first aspect above.

[0017] The personnel management method, system, and device based on facial recognition provided in this application embodiment have at least the following technical effects.

[0018] By deploying a face recognition-based personnel management method within a face recognition device—an integrated unit encompassing an electronic screen and processor for human-computer interaction—high hardware integration and simplified deployment are achieved, effectively reducing equipment costs and installation complexity. The face recognition device connects to face capture cameras and video surveillance cameras, acquiring face image data from these sources. This enhances the compatibility and acquisition capabilities of multi-source video data, providing richer and more reliable image sources for subsequent recognition. Face image data is filtered according to face image quality rules, selecting target face images that meet these rules and filtering out blurry, occluded, and other low-quality face data, thus improving the quality of input data and the accuracy of subsequent comparisons. The target face image data is compared with face features in a database to generate face recognition results, automating the process from image acquisition to identity verification, improving recognition efficiency and response speed. Personnel management is then implemented based on these face recognition results. By applying the identification results to business scenarios (such as stranger frequency statistics, personnel stay / departure duration analysis, and linkage between minor identification / verification / registration), proactive warnings of abnormal situations and intelligent monitoring of personnel dynamics can be provided, thereby improving personnel management. This addresses the question of how to improve personnel management within related technologies.

[0019] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of a personnel management method based on facial recognition; Figure 2 This is a flowchart illustrating step S102 according to an exemplary embodiment; Figure 3 This is a flowchart illustrating step S103 according to an exemplary embodiment; Figure 4 This is a system block diagram of a personnel management system based on facial recognition, according to one embodiment. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0022] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0023] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0024] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0025] It should be noted that the data collection, information extraction and processing steps involved in the embodiments of this application strictly comply with the requirements of relevant national laws and regulations (including but not limited to the Personal Information Protection Law of the People's Republic of China and the Data Security Law of the People's Republic of China).

[0026] Firstly, embodiments of this application provide a personnel management method based on facial recognition. Figure 1 This is a flowchart of a personnel management method based on facial recognition, such as... Figure 1 As shown, the method is deployed on a face recognition device, which is an integrated device that includes an electronic screen and a processor for human-computer interaction; the method includes: Step S101: Obtain facial image data captured by the face capture camera and video surveillance camera connected through the face recognition device.

[0027] Step S102: Filter the face image data according to the face image quality rules to obtain the target face image data that meets the face image quality rules.

[0028] Step S103: Compare the target face image data with the face features in the database to generate a face recognition result.

[0029] Step S104: Based on the facial recognition results, conduct personnel management.

[0030] In summary, this application provides a personnel management method based on face recognition. By deploying the face recognition-based personnel management method on a face recognition device, which is integrated into a single unit, the face recognition device integrates an electronic screen and processor for human-computer interaction, achieving high hardware integration and simplified deployment, effectively reducing equipment costs and installation complexity. Through face capture cameras and video surveillance cameras connected to the face recognition device, face image data collected by these cameras is acquired, enhancing the compatibility and acquisition capabilities of multi-source video data and providing richer and more reliable image sources for subsequent recognition. Face image data is filtered according to face image quality rules to obtain target face image data that meets the rules, filtering out low-quality face data such as blurry or occluded images, improving the quality of input data and the accuracy of subsequent comparisons. The target face image data is compared with face features in the database to generate face recognition results, realizing automated processing from image acquisition to identity verification, improving recognition efficiency and response speed. Personnel management is then carried out based on the face recognition results. By applying the identification results to business scenarios (such as stranger frequency statistics, personnel stay / departure duration analysis, and linkage between minor identification / verification / registration), proactive warnings of abnormal situations and intelligent monitoring of personnel dynamics can be provided, thereby improving personnel management. This addresses the question of how to improve personnel management within related technologies.

[0031] In one implementation, step S101 involves acquiring facial image data captured by a face capture camera and a video surveillance camera connected to a face recognition device. This specifically includes the following steps: Step S1011: Connect the video surveillance camera to the face recognition device and collect video stream data through the video surveillance camera; continuously track the face target in the video stream data according to the face capture algorithm, and obtain video face images that meet the preliminary face image quality rules according to the first preliminary face image quality rules; analyze the video face attribute information based on the video face image, and use the video face image and video face attribute information as the face image data collected by the video surveillance camera.

[0032] Optionally, a general-purpose video surveillance camera is responsible for acquiring and transmitting the raw video stream to the face recognition device. This device, acting as an edge computing node, runs a face capture algorithm to continuously detect and track facial targets in the video stream, and simultaneously performs real-time filtering based on a first preliminary face image quality rule (primarily evaluating the rate of change in face bounding box size during tracking). Through this process, the system selectively outputs video face images that meet the initial quality standards for each tracked target, extracts their attribute information (such as gender and age group), and finally binds the two together to form standardized face image data acquired from the video stream, suitable for subsequent deep processing.

[0033] Wherein: According to the face capture algorithm, the face target in the video stream data is continuously tracked, and a video face image conforming to the first preliminary face image quality rule is obtained according to the first preliminary face image quality rule, including: Continuous detection and tracking of target faces in video stream sequences; During the tracking process, based on the first preliminary face image quality rules, the face image is evaluated in real time until a preset trigger event occurs. The attributes used for quality evaluation include the rate of change of the face detection bounding box size. The trigger event includes at least one of the following: The target face leaves the video frame; The duration of continuous tracking of the target face exceeds the preset tracking timeout threshold; The number of video face images that meet the first preliminary face image quality rules exceeds the preset number; In response to the instruction of receiving the trigger event, the video face image that conforms to the first preliminary face image quality rules and the corresponding face attribute information are used as the video face image.

[0034] Optionally, during video stream face capture, each facial target in the video stream sequence is continuously detected, and independent, continuous face tracking trajectories are performed across frames or frame by frame. Throughout the tracking process, the target face image appearing in each frame is evaluated in real time based on a first preliminary face image quality rule. The core indicator of this evaluation is the inter-frame change rate of the face detection box size, used to quantify the scale stability of the target in the image and exclude drastic image scaling caused by the target suddenly accelerating towards or away from the camera. This evaluation process continues until any preset trigger event is met: 1) the target leaves the monitoring screen, and the tracking ends naturally; 2) the continuous tracking duration of the target reaches a preset timeout threshold to avoid indefinite tracking of the lingering target; 3) the number of images that have been accumulated from the target and meet the quality rules reaches a preset upper limit to prevent data redundancy. When any trigger event is detected, the system immediately responds to the instruction, no longer waiting for a better image, but instead using the current video face image that has passed the real-time quality evaluation and meets the first preliminary rule (i.e., the most scale-stable image) and its corresponding face attribute information as the final output to complete the capture task. The inter-frame change rate of the face detection box size is calculated as: (current frame face detection box size - previous frame face detection box size) / previous frame face detection box size.

[0035] For example, the scale stability of target motion is quantified by calculating the rate of change in the size of the face detection bounding box between consecutive video frames in real time. This rate of change is compared with a preset threshold. If it exceeds the threshold (e.g., a sudden scaling from a normal 10% increase to a drastic 36%), the frame is determined to be prone to motion blur due to the target suddenly accelerating closer or further away, and such unstable frames are automatically filtered out. The purpose is to ensure that the system only outputs clear face images with stable scale changes during motion, providing a high-quality image foundation for subsequent recognition.

[0036] Step S1011 performs preliminary screening using preliminary face image quality rules, which can effectively filter out invalid frames that are blurry or have distorted poses due to factors such as violent movement of the target, providing a stable and high-quality image source for subsequent processing and improving the input quality of the data processing chain.

[0037] Step S1012: By accessing the face capture camera, receive the captured face image collected by the face capture camera; according to the second preliminary face image quality rule, obtain the captured face image that conforms to the preliminary face image quality rule; analyze the captured face attribute information based on the captured face image, and use the captured face image and captured face attribute information as the face image data collected by the face capture camera.

[0038] Optionally, a face capture camera can be connected via the camera manufacturer's SDK or API interface. After connecting to a dedicated face capture camera, the system can receive single-frame captured face images actively reported by the camera in real time. For the characteristics of such static images, a second preliminary face image quality rule is applied for quality assessment. The core of this rule is to verify whether the size ratio of the face detection box is within a preset effective range. Specifically, the ratio of the height of the face detection box to the total image height, or the ratio of the area of ​​the face detection box to the total image area, is calculated, and it is determined whether this ratio is between a preset minimum threshold and a maximum threshold (for example, the face height should occupy 1 / 10 to 1 / 3 of the image height). If the ratio is lower than the minimum threshold, it indicates that the face is too small and cannot be used for recognition; if it is higher than the maximum threshold, it indicates that the face is too large, may have overflowed the image, or only contains a portion of the face. Only when the face size ratio is within the effective range does the capture pass the preliminary screening. Subsequently, the system analyzes the captured face attribute information based on this image, and finally outputs both as valid face image data.

[0039] It should be noted that the captured facial images include both facial screenshots and panoramic background images. Facial attribute information includes gender, age, ethnicity, skin color, hairstyle, hair color, 3D facial angle, blur or sharpness, brightness or darkness, facial completeness; as well as whether masks, glasses, or hats are worn; the degree of occlusion of the mouth, nose, and eyes, and the confidence values ​​of the above attributes.

[0040] Step S1012 uses a face size ratio as the second preliminary rule, which can quickly eliminate invalid captures caused by improper target distance (such as faces that are too small or too large) with extremely low computational overhead, ensuring the basic usability of the input data in terms of spatial scale.

[0041] In one implementation, Figure 2 This is a flowchart illustrating step S102 according to an exemplary embodiment, as follows: Figure 2 As shown, step S102 involves filtering face image data according to face image quality rules to obtain target face image data that conforms to the rules. Specifically, this includes the following steps: Step S1021: Perform a comprehensive quality assessment of the face image data according to preset face image quality rules; wherein the quality rules include at least: face 3D angle, image brightness, image sharpness, face integrity, face confidence, and the occlusion status of the eyes, mouth, and nose. The comprehensive quality assessment of the face image data according to the preset face image quality rules includes: Set an independent judgment threshold for each item in the quality rule; Determine whether the evaluation results of the face image data on all quality rule items meet the corresponding independent judgment thresholds; If so, the facial image data is deemed to have passed the comprehensive quality assessment; If not, the facial image data is deemed to have failed the comprehensive quality assessment.

[0042] Optionally, a comprehensive and rigorous quality assessment of the facial image data collected from the front end is performed using a set of multiple independent facial image quality rules. This set of rules includes at least: three-dimensional facial angles (yaw, pitch, and roll angles calculated by analyzing facial key points to assess whether the pose is controllable and frontal), image brightness (adjusting exposure based on the grayscale histogram of the overall image or the facial region), image sharpness (quantifying blur by calculating image gradients or frequency domain analysis), facial integrity (determining whether the face is completely within the image based on the position and boundary relationship of the face detection box in the image), face confidence (the confidence score output by the object detection model indicating that the region contains a face), and the occlusion status of the eyes, mouth, and nose (determining the presence of occlusions such as masks, glasses, and hands through local feature detection or classification models). The execution process is as follows: First, an independent, empirically or experimentally calibrated judgment threshold is preset for each of the above rules. Then, for the input facial image data, its quantitative evaluation results on all the above rules are calculated sequentially. Finally, the system performs a strict logical judgment: only when the evaluation results of the image on all rule items meet (i.e. reach or exceed) their respective independent judgment thresholds will the image be finally judged to have passed the comprehensive quality assessment and thus marked as target face image data that meets the quality requirements; if the evaluation result of any rule item fails to meet the standard, the image is judged as failing as a whole.

[0043] By using a multi-dimensional set of rules and setting independent thresholds for each rule for logical judgment, low-quality images with defects in any dimension such as pose, lighting, sharpness, integrity, or confidence can be filtered out, thus ensuring that only high-quality data enters the subsequent recognition process.

[0044] Step S1022: The evaluation results of the quality rule items are weighted according to the preset weights to obtain the comprehensive quality score of the face image data.

[0045] Optionally, firstly, a specific weight coefficient is assigned to each preset quality rule item (such as 3D angle, sharpness, occlusion status, etc.), representing the relative importance of this item in the overall evaluation. Then, the independent evaluation result of each rule item is converted into a standardized individual score. Finally, a single comprehensive quality score is calculated by multiplying each individual score by its corresponding weight coefficient and then summing all the products. This score comprehensively and quantitatively reflects the overall performance level of the face image across all key quality dimensions.

[0046] Step S1023: Compare the overall quality score with the preset overall quality threshold, and mark the face image data that is greater than or equal to the overall quality threshold as target face image data that conforms to the face image quality rules.

[0047] Optionally, one or more comprehensive quality thresholds are preset, representing the minimum comprehensive quality level. The comprehensive quality score of the current face image data is directly compared with the preset threshold: if the comprehensive quality score is greater than or equal to the threshold, the system determines that the image has reached an acceptable standard in all quality dimensions of weighted fusion, and then marks it as "target face image data that conforms to the face image quality rules" and allows it to flow into the subsequent face feature extraction and comparison process; conversely, if the comprehensive quality score is lower than the threshold, the overall quality of the image is determined to be insufficient, and it is excluded from the subsequent core processing process.

[0048] Step S102: Through multi-dimensional quality assessment, target face image data that meets the face image quality rules can be accurately selected.

[0049] In one implementation, Figure 3 This is a flowchart illustrating step S103 according to an exemplary embodiment, as follows: Figure 3 As shown, step S103 involves comparing the target face image data with the face features in the database to generate a face recognition result. Specifically, this includes the following steps: Step S1031: Extract facial features from the target face image data to generate a feature vector to be compared; Step S1032: Calculate the similarity between the feature vector to be compared and the N pre-stored face feature vectors in the database to obtain N similarity scores; Step S1033: Select the database personnel information corresponding to the highest similarity score from N similarity scores; Step S1034: Compare the highest score with a preset threshold, and determine the face recognition result based on the comparison result and the personnel information in the database. Specifically, this includes: If the highest score is greater than or equal to the preset threshold, the recognition is considered successful, and the database personnel information corresponding to the highest score is output as the face recognition result. If the highest score is less than the preset threshold, the recognition is deemed to have failed, and the target face image data is marked as stranger information as the face recognition result.

[0050] Optionally, after obtaining target face image data that meets the quality requirements, facial features are first extracted from the target face image data to generate a feature vector to be compared. Then, this feature vector is compared with N pre-stored face feature vectors in the database using a one-to-many (1:N) similarity calculation, resulting in N corresponding similarity scores. Next, the highest score is selected from these N scores, and the corresponding database personnel identity information is locked. Finally, a crucial decision step is executed: this highest score is compared with a preset judgment threshold that has been validated through extensive experiments. If the highest score is greater than or equal to the preset threshold, the system determines that the recognition is successful and officially outputs the database personnel information corresponding to the highest score as the final face recognition result; conversely, if the highest score is less than the preset threshold, the system determines that the recognition has failed, and the current target face image data is marked as "stranger" information, with this marking as the final result of this recognition.

[0051] Step S103, through feature extraction, similarity calculation, sorting, and threshold judgment, can quickly and objectively determine facial identity, avoiding the subjectivity and inefficiency of manual intervention, and significantly improving the processing speed and accuracy of the recognition business.

[0052] For example, hotels, online rental apartments, and public rental housing can issue warnings and remind non-residents (strangers) to register, or, based on the frequent and repeated appearances of non-residents (strangers) within a specific time frame, suspect them of involvement in prostitution, gambling, or drugs. This can be combined with prominent on-site reminders and data synchronization with an internal network platform to ensure data remains within the internal network, is secure, and is legal and compliant. Alternatively, in places where minors are prohibited or restricted, such as KTVs, bars, billiard halls, internet cafes, and card rooms, age-based warnings and registration can be implemented.

[0053] In one implementation, step S104, based on the facial recognition result, involves personnel control, including: Information on individuals identified as strangers is automatically archived, and their frequency of appearance in different time periods and regions is statistically analyzed. When the frequency of a stranger's appearance exceeds a preset threshold, an early warning mechanism is triggered. Based on continuous facial recognition records, the system calculates the duration of a specific person's stay in the monitored area. When the stay exceeds a preset time limit, an early warning mechanism is triggered. Based on the identified age attributes, minors are initially screened out, and at least one of the following operations is performed: Based on the manually entered name and ID number, obtain the ID information, and submit the ID information and a captured face photo or a face photo collected in cooperation by calling the trusted identity authentication interface provided by the public security department or a third party, and request the return of the verification result; Based on the connection of the ID card reader, the system automatically reads the ID card information and compares the portrait photo in the ID card information with the on-site captured face photo or the face photo collected in cooperation with the system's local algorithm to obtain the verification result; or it submits the ID card information and the on-site captured face photo or the face photo collected in cooperation with the system by calling the trusted identity authentication interface provided by the public security department or a third party, and requests the return of the verification result. After verification, the minor's identity information, facial photo, and recognition record will be linked and stored to complete the minor's information registration.

[0054] Optionally, firstly, each record of a stranger identified (including captured images, timestamps, and camera locations) is automatically added to a dedicated stranger database. The frequency of each stranger's appearance is calculated within time windows (e.g., 1 hour, 24 hours, 7 days, 30 days) and geographical areas. When the system detects that the frequency of a specific stranger's appearance in a set time period and area exceeds a preset threshold (e.g., the same stranger appearing on multiple nights within a week), an alert is automatically triggered, sending a notification to the front-end management system. Secondly, personnel loitering monitoring: Continuous trajectory tracking is performed on all identified personnel (including known individuals and strangers). By recording the timestamps of the first and last appearances of the same person in the monitored area, their loitering duration is automatically calculated. When the calculated duration exceeds a preset time limit for that area (e.g., loitering in a hotel corridor for more than 30 minutes), the system also triggers an alert. Thirdly, minor verification and registration: For individuals with an age attribute ≤18 years old, the system automatically initiates an enhanced verification process. The verification process offers two optional paths: 1) Calling a trusted identity authentication interface provided by the public security department or a third party, submitting a captured facial photo or a photo of a person collected in cooperation with the operator, along with the ID number and name manually entered by the operator, and requesting a verification result; 2) Connecting to an ID card reader via USB, automatically reading the ID card information, and using the system's local algorithm to compare the ID card photo with the captured facial photo or the photo of a person collected in cooperation with the operator, obtaining the verification result. Alternatively, the trusted identity authentication interface provided by the public security department or a third party can be called to submit the ID card information and the captured facial photo or the photo of a person collected in cooperation with the operator, requesting a verification result. After successful verification, the minor's identity information, the captured photo, the verification result, and the time and location information are associated to generate a registration record, which is then stored in the database, completing the registration process.

[0055] Traditional systems only record strangers, while this application automatically analyzes their behavioral patterns. For example, in entertainment venues, if a stranger repeatedly loiters outside different private rooms (exceeding the allowed frequency), the system will issue a real-time alert, reminding staff to pay attention, thereby effectively preventing suspicious behaviors such as "surveillance" and moving the security line forward.

[0056] By automating the calculation of stay duration, the system solves the problem of continuous monitoring that is difficult to achieve manually. For example, in the context of online rental rooms, the system can automatically detect when a visitor stays in a room for far longer than the booked time, triggering timely reminders and assisting managers in handling overstays or potential safety hazards, thus achieving refined management.

[0057] In internet cafes or bars, if a face resembling a minor is detected, not only will an alarm be triggered, but the business process will also be halted, requiring staff to verify the minor's identity. If the verification confirms the minor's identity, the minor will be registered; if a fake ID or someone else's ID is used, the verification process will be stopped. This protects minors while also mitigating legal risks for the operators.

[0058] In one implementation, after comparing the target face image data with face features in the database to generate a face recognition result, the method further includes: Personnel management includes searching, adding, modifying, and deleting personnel and their photos; re-extracting feature values ​​from personnel photos; and viewing detailed personnel information. The management of identification records includes retrieving and viewing details of identification records; identification records.

[0059] Optionally, by integrating personnel management and identification record management functions, full lifecycle management of the base database and identification logs is achieved. Operations such as adding, deleting, modifying, and querying personnel information, as well as feature re-extraction, ensure the accuracy and timeliness of the base database data. When a person's appearance changes or the initial photo quality is poor, re-extracting feature values ​​can directly improve the recognition rate and system adaptability of subsequent comparisons.

[0060] In one implementation, the method also includes: encrypting and de-identifying personnel data and captured data using a national cryptographic security chip built into or external to the facial recognition device, ensuring the security of data transmission and storage. It can also be connected to a regulatory or management platform to synchronize whitelisted and blacklisted personnel and report recognition records.

[0061] In summary, by integrating the facial recognition device into a single unit—combining the computing unit, touchscreen, and multiple network modules—the system achieves plug-and-play functionality, completely eliminating reliance on additional display devices, keyboards, mice, and dedicated computers. This significantly simplifies the deployment process, reduces hardware costs, and provides an intuitive and convenient front-end interactive experience. By supporting multiple external acquisition devices (such as facial capture cameras and regular surveillance video streams), the system can integrate heterogeneous data from multiple sources, achieving more comprehensive and accurate personnel perception and coverage of the monitored area. This provides high-quality, multi-dimensional factual evidence for subsequent in-depth analysis, laying the data foundation for precise management and control. By applying the recognition results to business scenarios (such as stranger frequency statistics, personnel stay / departure duration analysis, and linkage between minor identification / verification / registration), the system can proactively warn of abnormal situations, intelligently monitor personnel dynamics, and automatically complete compliance verification processes for specific groups (such as minors), significantly improving management efficiency, security levels, and compliance assurance capabilities.

[0062] Secondly, embodiments of this application provide a personnel management system based on facial recognition. Figure 4 This is a system block diagram illustrating a personnel management system based on facial recognition, according to one embodiment. Figure 4 As shown, the system includes a face image data module 410, a target face image data module 420, a face recognition module 430, and a personnel control module 440; wherein: The face image data module 410 is used to acquire face image data collected by the face capture camera and video surveillance camera connected through the all-in-one machine. The target face image data module 420 is used to filter face image data according to face image quality rules and obtain target face image data that conforms to the face image quality rules. The face recognition module 430 is used to compare the target face image data with the face features in the database to generate face recognition results.

[0063] The personnel management module 440 is used to manage personnel based on the facial recognition results.

[0064] In summary, this application provides a facial recognition-based personnel management system that, through a facial image data module, a target facial image data module, and a facial recognition module, automates the process from image acquisition to identity verification, improving recognition efficiency and response speed. It also creates files for strangers and, combined with age recognition, can be applied to scenarios requiring registration for entry and to places where minors are prohibited or restricted from entering. This addresses the problem of how to improve personnel management levels in related technologies.

[0065] It should be noted that the personnel management system based on facial recognition provided in this embodiment is used to implement the above-described embodiments, and details already described will not be repeated. As used above, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the above embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0066] Thirdly, embodiments of this application provide a personnel management device based on facial recognition, which may include: Electronic screens are used for human-computer interaction, avoiding the inconvenience of external keyboards and mice. Communication module, for user network communication; including 4G / 5G networks, Wi-Fi, and Ethernet; Cameras are used to collect facial information of targets for verification and / or registration; The USB interface connects to an external ID card reader, which is used to collect and verify the ID card information of the personnel. Combined with the security module of the facial recognition software, it can achieve secure protection for data transmission and storage. The system includes a memory and a processor, a program stored in the memory and executable on the processor, which, when executed by the processor, implements a face recognition-based personnel management method as described in the above embodiments.

[0067] In one embodiment, the facial recognition device is an integrated unit, further including a base, a wireless network, a wired network, a power interface, a power switch, and a memory card. An optional external ID card reader can be connected to collect and verify the ID card information of the person being recognized, and, in conjunction with the security module of the facial recognition software, to achieve secure data transmission and storage. An optional external national cryptographic security chip can be connected to achieve national cryptographic-level data encryption security.

[0068] This facial recognition device highly integrates an electronic screen, 4G / 5G network, camera, USB interface, and processor into a single unit, achieving a compact hardware architecture and coordinated operation of functional modules. This significantly reduces reliance on external devices and lowers deployment costs, while a built-in communication module ensures network connectivity in complex scenarios. The USB interface expands to include an ID card reader and integrates a security module, enabling dual verification of physical documents and biometrics. Combined with an optional national cryptographic security chip, data encryption capabilities are enhanced, effectively improving the reliability of the identity authentication process and the security of data transmission and storage, meeting compliance requirements for high-security scenarios. The integrated design, incorporating multiple network modules and security units, allows for rapid deployment and flexible networking in terminal scenarios such as hotel front desks and mobile policing, while ensuring the security of core data. This hardware-level design creates a facial recognition device that integrates convenience, security, and reliability.

[0069] Fourthly, embodiments of this application provide a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements a personnel management method based on face recognition provided in the first aspect.

[0070] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0071] In a possible implementation, the present invention can also be implemented as a program product comprising program code, which, when the program product is run on a terminal device, causes the terminal device to perform steps implementing the personnel management method based on face recognition provided in the first aspect.

[0072] The program code for executing the present invention can be written in any combination of one or more programming languages. The program code can be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on a remote device.

[0073] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0074] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A personnel management method based on face recognition, characterized in that, The method is deployed on a face recognition device, which is an integrated unit that includes an electronic screen and a processor for human-computer interaction; the method includes: The face image data collected by the face capture camera and video surveillance camera connected to the face recognition device is obtained. The face image data is filtered according to face image quality rules to obtain target face image data that conforms to the face image quality rules; The target face image data is compared with the face features in the database to generate a face recognition result; Personnel management is carried out based on the facial recognition results.

2. The personnel management method based on face recognition according to claim 1, characterized in that, The face capture camera and video surveillance camera connected through the face recognition device acquire face image data collected by the face capture camera and the video surveillance camera, including: The video surveillance camera is connected to the face recognition device, and video stream data is acquired through the video surveillance camera; according to the face capture algorithm, the face target in the video stream data is continuously tracked, and according to the first preliminary face image quality rule, a video face image conforming to the first preliminary face image quality rule is acquired; based on the video face image, video face attribute information is analyzed, and the video face image and the video face attribute information are used as the face image data acquired by the video surveillance camera; By accessing the face capture camera, the system receives captured face images collected by the face capture camera; according to the second preliminary face image quality rules, it obtains captured face images that conform to the second preliminary face image quality rules; based on the captured face images, it analyzes the captured face attribute information, and uses the captured face images and the captured face attribute information as face image data collected by the face capture camera.

3. The personnel management method based on face recognition according to claim 2, characterized in that, The step of continuously tracking facial targets in the video stream data and acquiring video facial images that conform to the first preliminary facial image quality rules includes: Continuous detection and tracking of target faces in video stream sequences; During the tracking process, the face image is evaluated in real time based on the first preliminary face image quality rules until a preset trigger event occurs. The quality evaluation is based on attributes including the rate of change of the face detection bounding box size. The trigger event includes at least one of the following: The target face leaves the video frame; The duration of continuous tracking of the target face exceeds a preset tracking timeout threshold; The number of video face images that meet the first preliminary face image quality rules exceeds the preset number; In response to the instruction of receiving the trigger event, the video face image that conforms to the first preliminary face image quality rules and the corresponding face attribute information are used as the video face image.

4. The personnel management method based on face recognition according to claim 1, characterized in that, The step of filtering the face image data according to face image quality rules to obtain target face image data that conforms to the face image quality rules includes: The face image data is comprehensively evaluated based on face image quality rules; wherein, the quality rules include at least: face three-dimensional angle, image brightness, image sharpness, face integrity, face confidence, and the occlusion status of eyes, mouth, and nose; The evaluation results of the quality rule items are weighted according to preset weights to obtain the comprehensive quality score of the face image data. The overall quality score is compared with a preset overall quality threshold, and face image data that is greater than or equal to the overall quality threshold is marked as target face image data that conforms to the face image quality rules.

5. The personnel management method based on face recognition according to claim 4, characterized in that, The comprehensive quality assessment of the face image data based on face image quality rules includes: Set an independent judgment threshold for each of the quality rule items; Determine whether the evaluation results of the face image data on all the quality rule items meet the corresponding independent judgment thresholds; If so, the facial image data is determined to have passed the comprehensive quality assessment; If not, the facial image data is determined to have failed the comprehensive quality assessment.

6. The personnel management method based on face recognition according to claim 1, characterized in that, The step of comparing the target face image data with face features in the database to generate a face recognition result includes: Facial features are extracted from the target face image data to generate a feature vector to be compared; The similarity between the feature vector to be compared and the N face feature vectors pre-stored in the database is calculated to obtain N similarity scores; Select the database personnel information corresponding to the highest similarity score from N similarity scores; The highest score is compared with a preset threshold, and the face recognition result is determined based on the comparison result and the personnel information in the database.

7. The personnel management method based on face recognition according to claim 6, characterized in that, The step of comparing the highest score with a preset threshold and determining the face recognition result based on the comparison result and the personnel information in the database includes: If the highest score is greater than or equal to the preset threshold, the recognition is determined to be successful, and the database personnel information corresponding to the highest score is output as the face recognition result. If the highest score is less than the preset threshold, the recognition is deemed to have failed, and the target face image data is marked as stranger information as the face recognition result. 8.The personnel management method based on face recognition of claim 1, wherein, Based on the facial recognition results, personnel management is implemented, including: Information on individuals identified as strangers is automatically archived, and their frequency of appearance in different time periods and regions is statistically analyzed. When the frequency of a stranger's appearance exceeds a preset threshold, an early warning mechanism is triggered. Based on continuous facial recognition records, the system calculates the duration of a specific person's stay in the monitored area. When the stay exceeds a preset time limit, an early warning mechanism is triggered. Based on the identified age attributes, minors are initially screened out, and at least one of the following operations is performed: Based on the manually entered name and ID number, obtain the ID information, and submit the ID information and a captured face photo or a face photo collected in cooperation by calling the trusted identity authentication interface provided by the public security department or a third party, and request the return of the verification result; Based on the connection of the ID card reader, the system automatically reads the ID card information and compares the portrait photo in the ID card information with the on-site captured face photo or the face photo collected in cooperation with the system's local algorithm to obtain the verification result; or it submits the ID card information and the on-site captured face photo or the face photo collected in cooperation with the system by calling the trusted identity authentication interface provided by the public security department or a third party, and requests the return of the verification result. After verification, the minor's identity information, facial photo, and recognition record will be linked and stored to complete the minor's information registration.

9. A personnel management system based on facial recognition, characterized in that, The system is used to execute the method according to any one of claims 1 to 8, and the system includes a face image data module, a target face image data module, a face recognition module, and a personnel control module; wherein: The face image data module is used to acquire face image data collected by the face capture camera and the video surveillance camera connected through the face recognition device. The target face image data module is used to filter the face image data according to face image quality rules to obtain target face image data that conforms to the face image quality rules. The face recognition module is used to compare the target face image data with face features in the database to generate a face recognition result; The personnel management module is used to manage personnel based on the facial recognition results.

10. A personnel management device based on facial recognition, characterized in that, include: Electronic screens are used for human-computer interaction; Communication module, user network communication; Cameras are used to collect facial information of targets for verification and / or registration; The USB interface connects to an external ID card reader, which is used to collect and verify the ID card information of the personnel. Combined with the security module of the facial recognition software, it can achieve secure protection for data transmission and storage. A memory and a processor, a program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements a face recognition-based personnel management method as described in any one of claims 1 to 8.

Images