Automatic Attendance System using a Group Image

Abstract

Marking employee attendance and keeping track of it is crucial to any organization, and this record is used as a benchmark for many other things. With an enormous number of employees working there, physically recording each employee's daily attendance is tiresome and time-consuming. Digital solutions based on RFID, biometrics and other technologies have been developed. However, they require expensive hardware in-frastructure and ongoing maintenance, which isn't cost-effective. We present a revolutionary real-time attendance marking and recording system that addresses all of these problems to meet the demand of the digital age for attendance marking without spoofing or unethical behaviors. Our solution is a user-friendly android application that allows users to take and upload a group snapshot of the people whose attendance needs to be noted. The application has a robust backend that effectively recognizes each user by detecting their faces, extracting those faces, creating face embeddings, and comparing them to those in the database. We have recorded the recognition time of 8.28 seconds. Our solution has the potential to be quickly adopted by any company with just a few modifications.