Malpractice Detection Using Machine Learning

Abstract

Cheating during exam is unfair behaviour of the students. This becomes disadvantage for other students who do not cheat and work hard to study. This project is meant to detect the suspicious activities of the students who intend to cheat during exams. This project is useful in educational institutions where in invigilators are not able to detect the cheating done by students in the exam hall. This project has two phases namely, training phase and the testing phase. The dataset is a pre-recorded video of students which is captured in real time. This system works well with live web camera input also. The algorithm used is DenseNet121.This system predicts three kinds of activities such as turning around, passing of paper and a normal activity differentiated by RGB colours.