Deep-learning based group-photo Attendance System using One Shot Learning
Aruna B. Bhat, Shivam Rustagi, Shivi R Purwaha, Shubhang Singhal · 2020 International Conference on Electronics and Sustainable Communication Systems (ICESC) · 2020
Smartphone based face attendance system is the need in today's digital world and age for universities and schools as smartphones with very good picture quality provided by its camera makes such systems really affordable and practical as far as implementation is concerned. Automatic attendance systems make this daily practice of marking attendance easy and highly efficient thus helping reduce time wasted during lectures for such administrative work. Face recognition-based attendance systems are one such biometric based attendance systems which are more secure and can evade multiple fake or proxy attendance practices easily. This paper proposes a novel face recognition-based attendance system which can work with group photo of a class providing us a list of present students. It uses one shot learning based face recognition technique for our system which can work for new users by providing only a single image of them thus making the system very robust and efficient. The proposed work presents a fully functional android app and backend system architecture which can easily be utilized by any university or school without requiring any expensive infrastructure setup. The proposed model achieves an accuracy of about 97% on LFW dataset and the proposed attendance system achieves an accuracy of 85% on a public student class photo dataset.