A Real Time Face Attendance Using Deep Learning

Abstract

One area of biometry technologies which is expanding quickly and being used extensively is face attendance in real time.Its uses are finding solutions for the industry, institutions, and law departments.The development of deep neural networks intended for all sets offace identification tasks, from recognition and preprocessing to image presentation and classification in verifying and finding answers, has become the primary focus of research due to the recent introduction of powerful, reasonably priced GPUs and the growth of massive face databases.The high processing expense of using Deep Convolutions Neural Networks (DCNN) and the need to strike a compromise between accuracy requirements and time and resource restrictions are the main reasons why realtime, exact image detection remains difficult despite recent advancements.Other important problems with face recognition include posture invariance, light, and occlusion, which significantly reduce accuracy in both deep neural networks and conventional handmade methods.It highlights areas that need further growth and improvement and offers a thorough analysis of both deep and shallow solutions .Inaddition to providing end users in business, government, and consumer with an insightful and critical viewpoint on currently available solutions, this review aims to support scientists\' and engineers\' research into new approaches of existing methodologies

Introduction

Traditional image recognition relied on manual feature extraction and statistical methods but faced challenges with varying lighting, angles, and obstructions. Deep learning, particularly convolutional neural networks (CNNs), improved recognition by automatically learning complex features, enhancing accuracy.

One practical system used OpenCV and Python on a Raspberry Pi with a camera to capture about 100 photos per student, detecting faces with Haar-like features and recognizing them using the LBPH technique.

Advanced models like DeepFace, FaceNet, and ArcFace further increased accuracy by capturing detailed facial features. Modern loss functions such as L2-constrained Softmax improve classification but require significant GPU resources and careful sample selection.

Many systems use CNN-based face recognition for attendance, converting video to images, detecting faces, and updating records, often storing attendance in Excel. Some mobile solutions allow remote attendance via a single camera.

The described methodology involves a sliding camera to capture classroom images, extracting facial embeddings with Deepface and FaceNet-512, and saving these for efficient face recognition and comparison.

Conclusion

Unauthorized users cannot alter the attendance records or report since access to them is secure.Externally authenticated users who might need to create policies based on the data stored in the system\'s database can access the webbased API.The autonomy module and visual perception are discussed in this study. It then goes on to describe the project\'s technologies and methodologies. .Even when subjects had a beard or other facial traits, or wore glasses, Haar-Cascades face detection performed incredibly well. The speed of the realtime video was adequate and there was no discernible frame lag.When all things are taken into account, LBPH and HaarCascades can be used as an affordable face recognition system.One example is a device that may automatically take attendance in class or identify known troublemakers in a store or mall and tell the owner to keep him vigilant.

References

[1] Baig, S., Geetadhari, K., Noor, M. A., &Sonkar, A. (2022). Face recognition- based attendancemanagement system by using machine learning. International Journal of Multi- disciplinary Research and Growth Evaluation, 3(3), pp. 1–4. https://doi.org-/10.54660/anfo.2022.3.3.1 [2] Satpute, R., Sontakke, S., Gondaliya, B., Sonawane, T. &Suryawanshi, K. (2020). Attendance Management System Using Face Recognition. biometrics, 7(05). [3] Ranjan, R., C. D. Castillo, and R. Chellappa, L2-constrained Softmax Loss for Discriminative Face Verification. 2017. [4] Essien, U., & Ansa, G. (2023). Attendance Based Systems for Face Recognition using a Camera Rail Approach. Researchers Journal of Science and Technology, 3(1), 14–28. Retrieved from https://www.rejost.com.ng/index.php/home/article/view/45 [5] Margaret Rouse. (2012). What is facial recognition? - Definition from WhatIs.com. [online] Available at: http://whatis.techtarget.com/definition/facial- recognition [Accessed 25 Mar. 2018 [6] Face Recognition Based Attendance System Nandhini R, Duraimurugan N. [7] https://www.google.co.in

Copyright

Copyright © 2025 Farooque Alam. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.