The use of unapproved methods to obtain an unfair edge in academic exams is known as academic dishonesty. This study proposes the use of deep learning, an advanced kind of machine learning, to apply human posture tracking to test-takers in order to detect any cheating. To help invigilators utilize human posture tracking, misconduct identification, and live video monitoring from the selected device, a web-based solution was developed. It also allowed invigilators to review archived evidence. At maximum computational load, the suggested system achieves almost 10 frames per second, which is practically real-time performance. It performed with an accuracy rate of 90%, an area under the receiver operating characteristic curve (AUROC) of 90%, and an F1-score of 89.64% when evaluated on a validated dataset.