FACETRACK AI: Smart Attendance with Deep Learning-based Face Recognition

Abstract

Many educational institutions have manual or semi-automated attendance systems. These methods are slow and they are susceptible to fraud including \"buddy punching\". In this paper we present FaceTrack AI, a smart attendance system based on face recognition using deep learning to address these problems. It uses MTCNN to detect faces and align the facial features and FaceNet to generate embedding of person and then they are compared using cosine similarity person. It can be accessed via web where you can mark the attendance and store it in a single place. The system is likely to be more efficient in real time, irrespective of lighting and facial changes, and especially for Indian face datasets. It\'s more accurate, automatic and efficient than the conventional method as discussed in recent research articles [1], [3] and [5]. It also takes into account security and scalability issues as described in [6] and [8].

Introduction

The text describes FaceTrack AI, a smart attendance management system that uses deep learning-based facial recognition to automate attendance in educational institutions and workplaces. Traditional attendance methods such as manual registers, RFID cards, and biometric systems are often slow, error-prone, require additional hardware, and are vulnerable to proxy attendance or “buddy punching.” To overcome these limitations, FaceTrack AI introduces a contactless, secure, and scalable attendance solution powered by artificial intelligence and computer vision.

The system uses MTCNN (Multi-task Cascaded Convolutional Networks) for accurate face detection and alignment, even under varying lighting conditions and facial poses. After detection, FaceNet generates facial embeddings—high-dimensional feature representations of faces—which are then matched using cosine similarity to identify individuals and record attendance with high accuracy and low false recognition rates. The system also supports real-time image or video frame processing, making it suitable for classrooms and office environments.

The literature review highlights how recent advances in CNNs and deep learning have significantly improved automated attendance systems compared to traditional methods. Previous studies explored techniques such as FaceNet, ResNet50, anti-spoofing, liveness detection, and timetable integration, but challenges such as scalability, security, dataset diversity, and real-world deployment still remain. FaceTrack AI addresses these gaps by combining robust facial recognition with a web-based management platform.

The comparative analysis shows that FaceTrack AI solves issues found in earlier systems:

• RFID systems suffer from card sharing and proxy attendance.
• Fingerprint systems are contact-based and raise hygiene concerns.
• Traditional face recognition systems struggle in real-life conditions.
• Systems without anti-spoofing are vulnerable to impersonation.
• Offline systems lack scalability and real-time access.
• Limited datasets reduce recognition accuracy for diverse populations.

The proposed system includes a web application that provides attendance tracking, data storage, reporting, user management, and real-time monitoring. The workflow begins with capturing images through a camera, detecting and aligning faces using MTCNN, extracting embeddings with FaceNet, and comparing them with stored embeddings in the database using cosine similarity. Once matched, attendance is automatically recorded.

Conclusion

In this paper, we have proposed an intelligent attendance system, FaceTrack AI, using deep learning-based face recognition. It takes attendance automatically in an efficient way using MTCNN to detect faces and FaceNet to build features for recognition recognition. It also provides a web-based platform to view and manage attendance. Our system addresses the limitations of conventional attendance systems by providing a contactless, fast and scalable system. Experimental results demonstrate it is practical for real-time with low latency and high accuracy. FaceTrack AI is a promising method for school attendance. Further research should focus on incorporating security measures (such as liveness detection), making the system less sensitive to variations in lighting conditions and making the system mobile.

References

[1] “Facial Recognition Attendance System With Deep Learning,” ScienceDirect, 2025. [Online]. [2] “Attendance System With Facial Recognition Using AI,” ResearchGate, 2025. [Online]. [3] “AI-Based Attendance System Using Face Recognition,” IJSRSET, 2025. [Online]. [4] “Face Detection and Recognition for Attendance in Classroom Using Machine Learning,” JOIV, 2024. [Online]. [5] “Enhanced Face Recognition Attendance System with Real-Time Tracking,” ScienceDirect, 2025. [Online]. [6] “Face Recognition-Based Attendance System with Anti-Spoofing,” ResearchGate, 2023. [Online]. [7] “AI-Based Facial Recognition Attendance System,” RSIS International, 2025. [Online]. [8] “Timetable Integrated Attendance System Using Face Recognition,” IJRASET, 2025. [Online]. [9] “Development of Attendance Monitoring System Using Facial Recognition,” ResearchGate, 2024. [Online]. [10] “Smart Attendance Management System Using Face Recognition,” IJFMR, 2024. [Online].

Copyright

Copyright © 2026 Ashish Kumar, Anubhav Somvanshi, Brijesh Shah, Ayush Singh. 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.