Face Attendance System

Abstract

Attendances are taken in every school and college. The convention attendance system consists of registers marked by teachers which may lead to human error and a lot of maintenance. The system proposed in this study is to deviate from such a traditional system and introduce a new approach to taking attendance using image processing. The system uses a Histogram of Oriented Gradients (HOG) and python libraries such as OpenCV, Dlib, and NumPy. As a human, the brain automatically recognizes a face instantly, but the computer is not capable of this high-level generalization. The system automatically starts taking snaps and then applies face detection and recognition technique to the given image the recognized students are marked as present and their attendance is updated with the corresponding time[2]. The working of the system is that it first looks at a picture and finds all the faces in it. second, it focuses on each face and can understand that even if a face is directed in a weird direction or under bad lighting. Third, the system comes up with 68 specific points called landmarks, that exist on every face example the top of the chin, the inner edge of each eyebrow, etc., and then picks out unique features of the face that can be used to tell it apart from other people. Finally, compare the unique features of that face to those already determined faces. Then the person can clock in into the system after the person clocks out, the system automatically transfers the data into an excel sheet.