Attendance Tracking by Facial Recognition

Document Type


Lead Author Type

CIS Masters Student


Dr. Yonglei Tao; taoy@gvsu.edu

Embargo Period



Current systems that are generally used for tracking attendance for online exams/courses is either manual or marked automatically by successful logins. The proposed system uses facial detection and recognition to mark the attendance. This can be further expanded to track employees, replace traditional paper attendance and so on. Facial recognition system also increases security apart from frauds, it ensures no accidental data is leaked to unauthorized persons and no human intervention is needed to monitor the attendance or registration.

The proposed system allows pre-registration of users so that their details can be stored in the database, it also stores several sample images of the user’s face. The program compares the user’s face accessed whose attendance is sought to be marked in the real time through the laptop/computer’s in-built camera. This is compared with the samples stored in the database and if a match is found, the attendance is marked automatically.

This technology is currently in use for high end security organizations, government offices, immigration services, etc. However, it is still not expanded to be used for general purposes as the skills to develop facial recognition systems were limited and expensive until recent developments in machine learning took place. The advancement in research for artificial neural networks (ANN) has provided us with the concept of deep learning which can be used interchangeably with ANN now-a-days. It dissects facial images into pixels and patterns of pixels. It then generates a general pattern for a particular image known as the histogram of Gradients and then runs the calculations for basic measurements of the features. Based on these calculations, algorithms like eigenfaces and Local Binary Patterns Histograms are used for further smarter processing.

This document is currently not available here.