Using Deep Learning to Detect Cheating on TCExam Platform Through Real-Time Facial Emotion Recognition

Abstract

Since the spread of the Covid-19 pandemic and its impact on the normal education process worldwide, including the conduct of exams, the education sector has had to shift from traditional in-person teaching to remote learning. This shift has led education authorities to confront the challenge of cheating in online exams and to seek reliable solutions to ensure fair exam conduct in an online environment. Our solution, based on machine learning techniques and artificial neural networks, aims to detect cheating cases in an online exam through facial emotion recognition exhibited by students during the online exam on the TCExam platform. To implement this approach, we need to prepare our own dataset. A well-trained, powerful, and effective system requires a large amount of data to achieve good results. In this article, we propose an architecture consisting of two parts: 1-Preparing the necessary dataset to train our system for generating fake images. We address the data issue by applying the Generative Adversarial Network (GAN) method. The results obtained subsequently will be the training data for our cheating detection system. 2-Feeding the facial emotion recognition techniques based on neural networks.