Publication Type : Conference Paper
Publisher : IEEE
Url : https://ieeexplore.ieee.org/document/8986206/
Keywords : cameras, face recognition, learning management systems, neural nets, smart phones
Year : 2019
Abstract : Attendance marking is a critical and time-consuming process in schools and colleges. Manual attendance marking is time consuming, so as attendance recording using biometrics. Attendance marking using face recognition is time saving when compared to conventional methods. Most of the existing face recognition systems which uses static cameras are expensive and have portability issues too. To overcome this above mentioned time and portability constraints we propose an attendance marking system based on face recognition. The proposed system implemented as an android application takes input from the smart phone camera to mark the attendance. It uses Facenet Resnet V1  convolutional neural network which was introduced by Google Inc, for face recognition. The attendance will be recorded in a learning management system(LMS) which serves as a back end application for the android application. After face recognition we are saving the attendance in our internal LMS system automatically. As per our analysis, we have noticed that the system works perfectly in a controlled scenario of 3-meter distance using a mobile camera device with a minimum face size of 108 × 108(Height × Width). In this controlled scenario our proposed methods achieves an accuracy over 90%.
Cite this Research Publication : N Pradeesh, VS Sreejesh Kumar, AS Anand, V Geetha Lekshmy, Shivsubramani Krishnamoorthy, Kamal Bijlani, Cost effective and reliable mobile solution for face recognition and authentication, 9th International Conference on Advances in Computing
and Communication, Pages 66 – 69, IEEE 2019 DOI: 10.1109/ICACC48162.2019.8986206