Vision Based Attendance System Using NVIDIA with Automated Notifications

Abstract

Attendance management in educational institutions is traditionally handled through manual or semi-automated systems that are often inefficient and prone to errors. This paper presents a smart attendance system based on Jetson Nano that utilizes facial recognition for real-time attendance marking. The system captures live video input, detects and recognizes faces using machine learning techniques, and automatically records attendancein acentralized database.Additionally, anautomated notification module informs students and administrative staff about attendance status through email or messaging services.The proposed solution reduces human intervention, enhances accuracy,andensurestransparencyinattendancemonitoring.

Introduction

This study presents an AI-powered smart attendance management system that uses facial recognition, edge computing, and automated notifications to improve the efficiency, accuracy, and transparency of attendance recording in educational institutions. Traditional attendance methods such as manual registers, RFID cards, and fingerprint systems are often time-consuming, prone to errors, vulnerable to proxy attendance, and may raise hygiene concerns due to physical contact.

Facial recognition offers a contactless and non-invasive alternative by identifying students through their unique facial features. Recent advances in Artificial Intelligence, particularly deep learning techniques such as Convolutional Neural Networks (CNNs) and face-embedding models, have significantly improved recognition accuracy even under challenging conditions such as varying lighting, facial expressions, and pose changes.

Many existing facial recognition systems rely on cloud computing, which can introduce latency, increase bandwidth consumption, and raise privacy concerns due to the transmission of sensitive biometric data. To address these issues, the proposed system uses edge computing with the NVIDIA Jetson Nano platform, enabling real-time face detection and recognition directly on the device. This reduces dependence on internet connectivity, lowers response time, and enhances data security by keeping biometric data locally stored.

The literature review highlights the evolution of attendance systems from manual methods to RFID, biometric authentication, traditional computer vision techniques, and modern deep learning-based facial recognition systems. While earlier methods suffered from limitations such as card misuse, hygiene issues, and poor performance under real-world conditions, deep learning models like FaceNet, DeepFace, and ArcFace have greatly improved recognition reliability. However, many existing systems still lack real-time feedback and notification mechanisms.

The proposed system addresses this gap by integrating automated notifications with attendance management. It consists of four main components: a camera module for capturing live video, a Jetson Nano device for face recognition, an attendance database for storing records, and a notification module that sends instant attendance confirmations to students and summary reports to faculty members.

The methodology involves collecting student facial images under different conditions, applying data augmentation techniques, detecting and recognizing faces using deep learning models, automatically marking attendance when confidence levels exceed predefined thresholds, and preventing duplicate entries. Notifications are then sent via email or messaging services.

The system is implemented using Python, OpenCV, TensorFlow/PyTorch, TensorRT optimization, and database technologies such as SQLite, MySQL, or cloud-based storage. Experimental results demonstrate high recognition accuracy, low latency, reduced dependence on internet connectivity, and improved transparency. Although factors such as poor lighting and facial occlusions can affect performance, these challenges can be mitigated through better dataset preparation and model optimization.

Conclusion

This paper introduced the design and implementation of a cost-effective smart attendance system utilizing facial recog- nition technology deployed on the NVIDIA Jetson Nano platform. The system, through the use of deep learning-based face embeddings and edge computing capabilities, effectively overcomes the shortcomings of the traditional attendance methodslikemanualrollcallsandRFID-basedsolutions.

References

[1] K. Mohammed et al., “Multimodal student attendance managementsystem,” Journal of King Saud University – Computer and InformationSciences, 2019. [2] M.WangandW.Deng,“Deepfacerecognition:Asurvey,”arXivpreprintarXiv:1804.06655, 2018. [3] A. Singh et al., “Smart attendance management system using facerecognition,” IJFMR, 2024. [4] S. K. Gupta et al., “CVUCAMS: Computer vision-based classroomattendance system,” ICSTCEE, 2020. [5] SoundaryaS.,AshwiniP.,RuchaW.,andGauravK.,“AReviewPaperonAttendance Management System Using Face Recognition,” InternationalJournal of Creative Research Thoughts (IJCRT), vol. 9, no. 11, 2021. [6] G.PrabhakarRaju,G.RohithReddy,M.Likhith,M.ManojKumar,and M. Hemanth, “Automated Face Recognition Attendance SystemUsingDeepLearning,”JournalofEmergingTechnologiesandInnovativeResearch (JETIR), vol. 11, no. 2, 2024. [7] A.J.Shepley,“DeepLearningforFaceRecognition:ACriticalAnalysis,” [8] CharlesDarwinUniversity,Australia,2019 [9] Pooja, J. Celia Grace, Rajalakshmi, and Rathi Jasna, “Face RecognitionAttendance,” International Journal of Novel Research and Development(IJNRD), vol. 9, no. 4, 2024. [10] V. Suresh, S. C. Dumpa, C. D. Vankayala, H. Aduri, and J. Rapa, “FacialRecognition Attendance System Using Python and OpenCV,” Journal ofSoftware Engineering and Simulation, vol. 5, no. 2, pp. 18–29, 2019 [11] T. Dhawle, U. Ukey, and R. Choudante, “Face Detection and Recog-nition using OpenCV and Python,” International Research Journal ofEngineering and Technology (IRJET), vol. 7, no. 10, 2020. [12] S. Singh and S. Graceline Jasmine, “Face Recognition System,” VelloreInstitute of Technology, Chennai, India [13] O.Akirke,A.Patange,D.Sonawane,C.Yenugwar,andS.S.Bhong,“ASurvey on Facial Recognition-Based Attendance Management System,”InternationalResearchJournalofModernizationinEngineeringTechnol-ogy and Science (IRJMETS), vol. 5, no. 11, 2023. [14] K.Yesugade,A.Pongade,S.Karad,D.Ingale,andS.Mahabare,“Face Detection and Recognition for Criminal Identification System,”International Journal of Creative Research Thoughts (IJCRT), vol. 12,no. 1, 2024. [15] N. R. Kurapati, “Face Recognition Based Attendance System UsingMachineLearningAlgorithms,”InternationalJournalforResearchTrendsand Innovation (IJRTI), vol. 9, no. 1, 2024. [16] A. Umalkar, S. S. Manhas, I. Chandiwala, and N. Bhagat, “Face Recog-nition Based Attendance System Using Real Time Data,” InternationalJournal of Creative Research Thoughts (IJCRT), vol. 11, no. 8, 2023

Copyright

Copyright © 2026 Kasalingam N, Gurusidda , Ganesh A P, Danesh M Kolavi, Pranith Bhandary. 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.