Non-inductive attendance management system and method based on face recognition

The face recognition-based contactless attendance management system utilizes AI cameras and artificial intelligence algorithms to achieve automated and intelligent management of employee attendance, solving the problems of time-consuming, labor-intensive, and inaccurate traditional attendance methods, and improving the accuracy and convenience of attendance.

CN118982334APending Publication Date: 2024-11-19COLORFUL GUIZHOU IMPRESSION NETWORK MEDIA CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411112281.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

Traditional attendance methods are time-consuming and labor-intensive, susceptible to environmental factors, and have low accuracy, especially manual sign-in, fingerprint recognition, and constrained facial recognition.

Method used

The system adopts a face recognition-based contactless attendance management system, which uses AI cameras, image processors and artificial intelligence algorithms for automated and intelligent management. Through face detection and recognition technology, combined with wireless communication and big data analysis, it achieves automated and accurate employee attendance.

Benefits of technology

It improves the accuracy and convenience of attendance tracking, enhances the efficiency of store employee management, and boasts high recognition accuracy and system stability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118982334A_ABST
    Figure CN118982334A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a non-inductive attendance management system based on face recognition, which comprises system management, face detection and recognition and hardware equipment, the hardware equipment is electrically connected with the face detection and recognition through data transmission, the face detection and recognition is electrically connected with the system management through data uplink, and the hardware equipment is electrically connected with the system management through data uplink. The system management is electrically connected with face detection and recognition through knowledge downlink, and the system management is internally and electrically connected with user management, service setting, application management, resource management, model deployment and feature library management; through a wireless communication technology, a face recognition technology and a big data analysis technology, automatic and intelligent management of employee attendance checking is realized, and the accuracy and convenience of attendance checking are improved. Meanwhile, the embodiment of the invention has relatively high identification accuracy and system stability, and can effectively improve the management efficiency of attendance checking of the store employees.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a non-contact attendance management system and method based on face recognition. Background Art

[0002] In the store employee attendance management, traditional attendance methods mainly include manual sign-in, fingerprint recognition, and constrained face recognition. These traditional attendance methods have many problems, such as manual sign-in is time-consuming and labor-intensive, fingerprint recognition may be affected by environmental factors (such as wet fingers, etc.), resulting in recognition failure, and constrained face recognition also has problems such as queuing. Therefore, how to improve the accuracy and convenience of attendance has become an urgent problem to be solved. Summary of the invention

[0003] The purpose of the present invention is to provide a non-contact attendance management system and method based on face recognition to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions: A non-contact attendance management system based on face recognition comprises system management, face detection and recognition and hardware equipment, wherein the hardware equipment is electrically connected to the face detection and recognition through data transmission, the face detection and recognition is electrically connected to the system management through data uplink, the system management is electrically connected to the face detection and recognition through knowledge downlink, and user management, service setting, application management, resource management, model deployment and feature library management are electrically connected inside the system management.

[0005] As a preferred solution of the present invention, the face detection and recognition are electrically connected internally with data collection, feature extraction, human body detection, face detection, face registration, data synchronization, familiarity calculation, face recognition and configuration synchronization.

[0006] As a preferred solution of the present invention, a smart camera, a networking device and a display device are respectively arranged inside the hardware device, and the smart camera, the networking device and the display device are all electrically connected to the hardware device.

[0007] A non-contact attendance management method based on face recognition includes human body collection, face detection, recognition stage and system application, wherein the human body collection comprises the following steps: Through AI cameras, it is possible to capture human body images through image recognition and analysis technology. AI cameras are usually composed of cameras, image processors and artificial intelligence algorithms. They can capture human body information through high-definition images.

[0008] As a preferred solution of the present invention, the image processor is responsible for converting the image captured by the camera into a digital signal for subsequent processing, and the artificial intelligence algorithm analyzes and processes the captured image to extract the human body image.

[0009] As a preferred solution of the present invention, the face detection includes two steps: first and second: First, you need to slide windows of different sizes and positions in the image, and then determine whether there is a face in the window, and use the face key point detection model in a complex environment based on the yolov5_adaloss enhanced loss function.

[0010] As a preferred solution of the present invention, the second step includes using the predicted key points to obtain a similarity transformation matrix M with preset standard frontal reference key points, and using the transformation matrix M to perform an affine transformation on the face image to complete face correction.

[0011] As a preferred solution of the present invention, the recognition stage includes feature vector extraction and distance matching, and the feature vector extraction includes the following steps: Good features need to have similarity preservation, that is, in the feature space, the feature similarity of images with high similarity is relatively high, and the feature similarity of image pairs with low similarity is relatively small. The detected face returns a 128-dimensional face feature vector through the interface of the face feature extraction model. Based on the 128-dimensional features of multiple face pictures of a single person at multiple angles, distances, and lighting conditions, re-analysis is performed to form unique feature data for Euclidean distance calculation.

[0012] As a preferred solution of the present invention, the distance matching includes the following steps: after obtaining the feature vector, the Euclidean distance and the local face feature vector can be used for matching, and the label of the sample is returned using the nearest neighbor classifier.

[0013] As a preferred solution of the present invention, the system application includes the following steps: For multiple pictures collected by individuals, they can also be taken with different expressions, postures, makeup, lighting, and blurred conditions for registration. For employees in stores, factories, and office buildings, employee facial identification is performed regularly, the time of entering and leaving the office area is recorded, and the data on entry and exit of personnel is pushed to the human resources system to make statistics on on-the-job and off-the-job, and working hours.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. In the present invention, wireless communication technology, face recognition technology and big data analysis technology are used to realize the automated and intelligent management of employee attendance, thereby improving the accuracy and convenience of attendance. At the same time, the embodiments of the present invention have a high recognition accuracy rate and system stability, and can effectively improve the management efficiency of store employee attendance.

[0015] 2. In the present invention, statistics are collected on the on-the-job work status of each employee in each store. When an employee enters or leaves a store area, facial image information is obtained through a capture device and facial recognition is performed. When it is determined that the employee is an employee of the store, the employee's arrival / departure record is made and the record information is returned to the human resources system for statistics. Based on the employee's arrival / departure statistical record table in the human resources system, an analysis of the corresponding employee's working hours is obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is an architecture diagram of the cloud-edge collaborative intelligent recognition system based on the present invention; Figure 2 This is a face recognition flow chart of the present invention; Figure 3 This is the architecture diagram of the contactless attendance management system based on face recognition of the present invention.

[0017] In the figure: 1. System management; 101. User management; 102. Service settings; 103. Application management; 104. Resource management; 105. Model deployment; 106. Feature library management; 2. Face detection and recognition; 201. Data collection; 202. Feature extraction; 203. Human body detection; 204. Face detection; 205. Face registration; 206. Data synchronization; 207. Familiarity calculation; 208. Face recognition; 209. Configuration synchronization; 3. Hardware equipment; 301. Smart camera; 302. Networking equipment; 303. Display equipment. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0019] For examples, see Figure 1-3 , the present invention provides a technical solution: A non-sensing attendance management system based on face recognition, including system management 1, face detection and recognition 2 and hardware device 3, the hardware device 3 is electrically connected to the face detection and recognition 2 through data transmission, the face detection and recognition 2 is electrically connected to the system management 1 through data uplink, the system management 1 is electrically connected to the face detection and recognition 2 through knowledge downlink, and the system management 1 is electrically connected to user management 101, service setting 102, application management 103, resource management 104, model deployment 105 and feature The database management 106 and the face detection and recognition 2 are electrically connected with data collection 201, feature extraction 202, human body detection 203, face detection 204, face registration 205, data synchronization 206, familiarity calculation 207, face recognition 208 and configuration synchronization 209 respectively. The hardware device 3 is internally provided with a smart camera 301, a networking device 302 and a display device 303, and the smart camera 301, the networking device 302 and the display device 303 are all electrically connected to the hardware device 3.

[0020] Among them, through wireless communication technology, face recognition technology and big data analysis technology, the automatic and intelligent management of employee attendance is realized, which improves the accuracy and convenience of attendance. At the same time, the embodiment of the present invention has a high recognition accuracy rate and system stability, which can effectively improve the management efficiency of store employee attendance.

[0021] A non-contact attendance management method based on face recognition includes human body collection, face detection, recognition stage and system application. The steps of human body collection are as follows: Through AI cameras, it is possible to capture human images through image recognition and analysis technology. AI cameras are usually composed of cameras, image processors, and artificial intelligence algorithms. They can capture human information through high-definition images. The image processor is responsible for converting the images collected by the camera into digital signals for subsequent processing. The artificial intelligence algorithm analyzes and processes the collected images to extract human images. Face detection consists of two steps: first and second: First, it is necessary to slide windows of different sizes and positions in the image, and then determine whether there is a face in the window. The face key point detection model in a complex environment based on the yolov5_adaloss enhanced loss function is used. Secondly, the predicted key points are used to obtain the similarity transformation matrix M with the preset standard frontal reference key points. The transformation matrix M is used to perform affine transformation on the face image to complete face correction.

[0022] The recognition stage includes feature vector extraction and distance matching. Feature vector extraction includes the following steps: Good features need to have similarity preservation, that is, in the feature space, images with high similarity have relatively high feature similarity and are close in distance, while images with low similarity have relatively small feature similarity and are far in distance. The detected face returns a 128-dimensional face feature vector through the interface of the face feature extraction model. The 128-dimensional features of multiple face images of a single person at multiple angles, distances, and lighting conditions are re-analyzed to form unique feature data for Euclidean distance calculation.

[0023] Distance matching includes the following steps: after obtaining the feature vector, the Euclidean distance can be used to match the local face feature vector, and the label of the sample can be returned using the nearest neighbor classifier.

[0024] System application includes the following steps: For multiple pictures collected by individuals, they can also be taken with different expressions, postures, makeup, lighting, and blurred conditions for registration. For employees in stores, factories, and office buildings, employee facial identification is performed regularly, the time of entering and leaving the office area is recorded, and the data on entry and exit of personnel is pushed to the human resources system to make statistics on on-the-job and off-the-job, and working hours.

[0025] The on-the-job work status of each employee in each store is counted. When an employee enters or leaves a store area, facial image information is obtained through a capture device and facial recognition is performed. If the employee is determined to be an employee of the store, the employee's arrival / departure record is made and the record information is returned to the human resources system for statistics. Based on the employee's arrival / departure statistical record table in the human resources system, an analysis of the corresponding employee's working hours is obtained.

[0026] Workflow of the present invention: When the non-sensing attendance management system and method based on face recognition designed by this scheme is working, firstly, the AI ​​camera is a camera integrated with artificial intelligence technology, which can realize the capture of human body images through image recognition and analysis technology, and it can capture human body information through high-definition images. The image processor is responsible for converting the image collected by the camera into a digital signal for subsequent processing. The artificial intelligence algorithm analyzes and processes the collected images, extracts human body images, slides in the image with windows of different sizes and positions, and then determines whether there is a face in the window, and uses the face key point detection model in a complex environment based on the yolov5_adaloss enhanced loss function, and uses the predicted key points to obtain the similarity transformation matrix M with the preset standard front face reference key points. Using the transformation matrix M, the face image is affinely transformed to complete face correction, realize face alignment transformation, and solve the face detection problem in complex scenes. Feature extraction is a key link in image recognition. Its function is to convert the input picture into a feature vector of fixed dimension for subsequent vector retrieval. Good features need to have similarity retention, that is, in the feature space, the feature similarity of images with high similarity is relatively high (the distance is relatively close), and the feature similarity of pairs of images with low similarity is relatively small (the distance is relatively far). The detected face returns a 128-dimensional face feature vector through the interface of the face feature extraction model. Based on the 128-dimensional features of multiple face pictures of a single person at multiple angles, multiple distances, and lighting, re-analysis is performed to form unique feature data for Euclidean distance calculation. After obtaining the feature vector, the Euclidean distance can be used to match the local face feature vector, and the label of the sample can be returned using the nearest neighbor classifier. For multiple pictures collected by individuals, they can also be taken from different expressions, postures, makeup, lighting, and blurry conditions for registration. For employees in stores, factories, and office buildings, employee face identification is performed regularly, the time of entering and leaving the office area is recorded, and the data of personnel entering and leaving the office is pushed to the human resources system to make statistics on on-the-job and off-the-job, and working hours.

[0027] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A non-contact attendance management system based on face recognition, comprising system management (1), face detection and recognition (2) and hardware equipment (3), characterized in that: The hardware device (3) is electrically connected to the face detection and recognition (2) via data transmission, the face detection and recognition (2) is electrically connected to the system management (1) via data uplink, the system management (1) is electrically connected to the face detection and recognition (2) via knowledge downlink, and the system management (1) is electrically connected to user management (101), service settings (102), application management (103), resource management (104), model deployment (105) and feature library management (106) respectively.

2. According to claim 1, a non-contact attendance management system based on face recognition is characterized in that: The face detection and recognition (2) is electrically connected internally with data collection (201), feature extraction (202), human body detection (203), face detection (204), face registration (205), data synchronization (206), familiarity calculation (207), face recognition (208) and configuration synchronization (209).

3. According to claim 1, a non-contact attendance management system based on face recognition is characterized in that: The hardware device (3) is provided with a smart camera (301), a networking device (302) and a display device (303) respectively, and the smart camera (301), the networking device (302) and the display device (303) are all electrically connected to the hardware device (3).

4. A non-contact attendance management method based on face recognition, applicable to a non-contact attendance management system based on face recognition as described in any one of claims 1 to 3 above, including human body collection, face detection, recognition stage and system application, characterized in that: The human body collection steps are as follows: Through AI cameras, it is possible to capture human body images through image recognition and analysis technology. AI cameras are usually composed of cameras, image processors and artificial intelligence algorithms. They can capture human body information through high-definition images.

5. According to claim 4, a non-contact attendance management method based on face recognition is characterized in that: The image processor is responsible for converting the images captured by the camera into digital signals for subsequent processing, and the artificial intelligence algorithm analyzes and processes the captured images to extract human body images.

6. According to claim 4, a non-contact attendance management method based on face recognition is characterized in that: The face detection includes two steps: first and second: First, you need to slide windows of different sizes and positions in the image, and then determine whether there is a face in the window, and use the face key point detection model in a complex environment based on the yolov5_adaloss enhanced loss function.

7. The method for non-contact attendance management based on face recognition according to claim 6 is characterized in that: The method secondly includes using the predicted key points to obtain a similarity transformation matrix M with preset standard frontal face reference key points, and using the transformation matrix M to perform an affine transformation on the face image to complete face correction.

8. According to claim 4, a non-contact attendance management method based on face recognition is characterized in that: The recognition stage includes feature vector extraction and distance matching, and the feature vector extraction includes the following steps: Good features need to maintain similarity, that is, in the feature space, images with high similarity have relatively high feature similarity (closer distance), and images with low similarity have relatively small feature similarity (farther distance). The detected face returns a 128-dimensional face feature vector through the interface of the face feature extraction model. The 128-dimensional features of multiple face images of a single person at multiple angles, distances, and lighting conditions are re-analyzed to form unique feature data for Euclidean distance calculation.

9. The method for non-contact attendance management based on face recognition according to claim 8 is characterized in that: The distance matching comprises the following steps: after acquiring the feature vector, the Euclidean distance can be used to match the local face feature vector, and the label of the sample can be returned using the nearest neighbor classifier.

10. The method for non-contact attendance management based on face recognition according to claim 4 is characterized in that: The system application comprises the following steps: For multiple pictures collected by individuals, they can also be taken with different expressions, postures, makeup, lighting, and blurred conditions for registration. For employees in stores, factories, and office buildings, employee facial identification is performed regularly, the time of entering and leaving the office area is recorded, and the data on entry and exit of personnel is pushed to the human resources system to make statistics on on-the-job and off-the-job, and working hours.