A cheating detection system in online exams through real-time facial emotion recognition of students

Since the onset of the Covid-19 pandemic and its widespread impact on global educational processes, including exam administration, the education sector has been compelled to transition from traditional in-person teaching to remote learning. This shift has exposed educational authorities to the challenge of cheating in online exams, prompting them to seek robust solutions to ensure exam integrity in a virtual environment. Our solution, grounded in machine learning techniques and artificial neural networks, aims to detect instances of cheating during online exams by analyzing students' emotional expressions during online assessments. Implementing this approach requires the creation of a specific and tailored dataset. A high-performing and effective system necessitates a substantial amount of data to achieve satisfactory results. In this article, we propose an architecture that facilitates both the preparation of the dataset required to train the system by applying the Generative Adversarial Network (GAN) method through the enhanced version of StyleGAN2 ADA to overcome challenges associated with the limited availability of data during training. The outcomes of this stage will serve as training data for the fraud detection system, which employs facial emotion recognition techniques based on neural networks to analyze students' facial expressions in real time during online exams.