REAL-TIME STUDENT ATTENDANCE CAPTURING WITH FACIAL RECOGNITION FROM SURVEILLANCE VIDEO USING CONVOLUTIONAL NEURAL NETWORK

Abstract

REAL-TIME STUDENT ATTENDANCE CAPTURING WITH FACIAL RECOGNITION FROM SURVEILLANCE VIDEO USING CONVOLUTIONAL NEURAL NETWORK Face recognition has been an active and vital topic among the computer vision community playing an increasingly important role in modern life and has been widely used in residential security, face authentication, and criminal investigation. Most techniques process visual data and search for general patterns in human faces. Ever since computers were developed, scientists and engineers thought of Artificial intelligence systems that are mentally and/or physically equivalent to humans. In the past decades, the increase of generally available computational power provided a helping hand for developing fast learning machines, whereas the internet supplied an enormous amount of data for training. These two developments boosted the research on smart self-learning systems, with neural networks among the most promising techniques. This students’ attendance capturing system uses a Convolutional Neural Network (CNN) algorithm a special type of deep learning Neural Network that is mainly used for image classification tasks. The image recognition performance of the system is improved with Convolutional Neural Network architecture that solves the facial image-related problems that certain varying illumination, poses, occlusion, and facial expressions. CNN shows an important improvement in attendance management technology, offering a solution that addresses societal needs for efficient and accurate identification. With CNN, this attendance capturing system shows the prospective to transform traditional methods of attendance tracking in educational settings. By overcoming limitations related to conventional techniques, such as manual data entry and inaccuracies in identification, this system offers a modernized approach to attendance management. FIG.1

Specification

Description:REAL-TIME STUDENT ATTENDANCE CAPTURING WITH FACIAL RECOGNITION FROM SURVEILLANCE VIDEO USING CONVOLUTIONAL NEURAL NETWORK
Technical Field
[0001] The embodiments herein generally relate to a method for real-time student attendance capturing with facial recognition from surveillance video using convolutional neural network.
Description of the Related Art
[0002] The technology plays a major role in all sectors. With the rapid technology, the mind-set of the people is adapted with the word instant and they expect the products with such a solution. Face recognition is one of the most important technologies that is used in many sectors to help the crime branch to identify the criminals, in education for biometric attendance, in traffic people violating traffic rules etc. This face recognition one such a greatest technology that helps and solves all the societal needs of the people. The integration of face recognition technology into the educational sector, helps in identifying the students and enhance the educational process effectively.
[0003] The biometric attendance systems with face recognition technology restructure attendance tracking processes in educational institutions. By accurately identifying the students, this system reduces the need for manual attendance registers, decreasing administrative loads and enhancing overall efficiency. Immediate verification of identities confirms accountability and transparency in academic learning environments. The implementation of a video-based face recognition attendance system provides several benefits to educational institutions in attendance management. With innovative technologies such as Machine Learning and Facial Recognition, this system renovate classroom processes and improves the overall learning experience for students and educators similarly.


SUMMARY
[0004] In real-time student attendance capturing with facial recognition from surveillance video using convolutional neural network. To streamline the attendance tracking process in educational institutions, reducing administrative loads related with manual attendance recording and data entry. To increase the accuracy and reliability of attendance records, reducing errors and differences in student attendance tracking.
[0005] To increase the accuracy and reliability of attendance records, reducing errors and differences in student attendance tracking. To improve the overall student experience by reducing disturbances in class attendance attendance records, reducing errors and differences in student attendance tracking. To improve the overall student experience by reducing disturbances in class attendance, reducing wait times for attendance recording, and promoting a continuous learning environment. To enhance resource utilization within educational institutions, allowing administrators to allocate time and resources more efficiently.
[0006] Currently, the educational institutions are using fingerprint biometric systems for attendance tracking purposes. This systems usually include the use of fingerprint scanners placed at entrances of the classrooms, where students are required to place their finger on the scanner to record their attendance.
[0007] Existing system is built up of outdated technology it works only when the user approaches the facial recognition then it recognizes the face of the user.
[0008] Also, students' attendance system with fingerprint recognition poses a problem when the student's finger is damaged.
[0009] The existing system makes the students' to wait for a longer time to give attendance leading to disturbance of classes.
[0010] Some persons may misuse this technology by just giving the facial recognition attendance and try to skip their work.
[0011] It consumes lot of user's time.
[0012] Accuracy is less.
[0013] The proposed system uses Convolutional Neural Network approach which helps to identify the captured face image from the video.
[0014] The web cam helps to detect the user faces automatically in 360 degrees.
[0015] The wastage time of the user is reduced.
[0016] Using Convolutional Neural Network algorithm, better results can be achieved when compared to previous research work.
[0017] This technology works at each and every unit of time by the given time format.
[0018] And also, it gives the information about attendance of the user in a detailed manner with absolute timings (in/out).

BRIEF DESCRIPTION OF THE DRAWINGS
[0019] The embodiments herein will be better understood from the following detailed description with reference to the drawings, in which:
[0020] FIG. 1 illustrates a method for real-time student attendance capturing with facial recognition from surveillance video using convolutional neural network according to an embodiment herein; and
[0021] FIG. 2 illustrates a method proposed for the admin gathers all the student information corresponding to a particular class, capturing his/her face images, train them and store it forming a dataset to an embodiment herein.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
[0022] The embodiments herein and the real-time student attendance tracking using facial recognition leverages modern computer vision techniques, particularly convolutional neural networks, to identify and record the presence of students by analyzing video feeds from surveillance cameras. This approach offers an automated, efficient, and non-intrusive way to monitor attendance in educational institutions.
[0023] FIG. 1 illustrates a method for real-time student attendance capturing with facial recognition from surveillance video using convolutional neural network Data collection and Training: This work started with the collection of student information such as personal details and facial images, which are stored in a database. The Machine Learning process analyses and learns the facial features of enrolled students, allowing correct identification and verification. The dataset is formed by collecting the faces in different poses.
[0024] Video-based Image capture: For capturing student attendance, the system utilizes video-based image capture techniques. During class sessions, a video recording device captures footage of students present in the classroom. Afterwards, the system extracts individual frames from the video, separating facial images of each student for further processing. This video-based approach enables real-time attendance tracking and reduces manual intervention.
[0025] Face Extraction and Matching: The system used Convolutional Neural Network algorithm to match the faces with the pre-trained images stored in the database, once the facial images are extracted from the video frames. The system compares facial features such as distance between eyes, nose shape, and facial contours to recognise possible matches. If a match is found between the captured image and the stored student data, the system marks the student's attendance consequently.
[0026] Attendance Reporting: Once the attendance tracking process completed, a complete report is generated as per the requirement of the admin. This report provides a detailed understandings into student attendance that includes attendance percentages, late arrivals, and absences facilitating decision-making and supports academic planning initiatives
[0027] FIG. 2 illustrates a method proposed for the admin gathers all the student information corresponding to a particular class, capturing his/her face images, train them and store it forming a dataset according to an embodiment herein. In some embodiments, A convolutional neural network, trained on a dataset of student faces, is employed to identify individual students. The CNN architecture typically includes multiple convolutional layers for feature extraction, followed by pooling layers and fully connected layers for classification. The model outputs the identity of students by matching detected faces against a stored database of registered student profiles. Popular CNN architectures for this task include VGGNet, ResNet, or custom lightweight models optimized for real-time performance.
S.No. Action Input Expected Output Actual Output Test Result
1 Capture Student face Images Students' face Images are captured and stored Images are captured and stored Pass
2 Train the image dataset Store the images of a face Create histogram and store Histogram is created and values are stored Pass
3 Face recognition A live stream of a students' face Name of detected student is displayed on the screen Name of detected student is displayed on the screen Pass
4 Update attendance for multiple students' at once Multiple faces from a live video stream Update attendance for all faces detected Attendance is updated only for a single face Fail
5 Detect more than 7 faces 7 students facing the camera Detect all 7 students faces present in front of camera Only 5 faces detected at a time Fail

, Claims:1. A method for a real-time student attendance capturing with facial recognition from surveillance video using convolutional neural network.
2. Video input module: continuously captures video from classroom surveillance cameras and processes the footage in real time. Face detection mechanism: utilizes deep learning-based face detection algorithms to locate and extract faces from each video frame.
3. Facial Recognition Module: Implements a pre-trained convolutional neural network designed to identify and match student faces against a stored database, ensuring accurate recognition and distinguishing between registered and unregistered individuals.
4. Attendance Logging System: Records attendance data, including time-stamped entries, and updates attendance records without redundancy for continuous or repeated detections.
5. Data Management and Security Layer: Secures student identity and attendance records by adhering to data privacy standards and employing encrypted storage solutions.

Documents