Table 1.

List of Past References, including methodology, dataset, techniques, and results.

Ref.	Dataset	Methodology	Results
[1]	Detail of the student cheating dataset: - Number of video sequences: 38 - Each class has an average number of images: 1660. - Testing images for each class: 510. - Unique subject recorded: Eight	Feature Extraction, ML, MSER Features, SURF, HOG, Robust Features.	SURF accuracy: 92% MSER Accuracy: 89% HOG Accuracy: 87%
[12]	Total Images of the Dataset: 7600.	3D CNN, Deep-Learning, Cheating in Exams, ML, Gesture Recognition Model, Object Detection.	- Lstm Model Accuracy: 0.77 - RNN Model Accuracy: 0.73 - 3DCNN Model Accuracy: 0.94.
[16]	- Online Participants from 5 different countries: 31 Participants. - Where Twenty-one students answered 379 questions.	Usability Online Examination,	- Students’ answers were 99.3% correct.
[24]	Students took part in the Exams: 104 students. - For Lab exams: six groups of up to twenty-two members were included.	Learning Management System, Learning Analytics. Python Tool	- Around twenty-three percent of students failed the test. - Whereas the previous result showed that forty-five percent of students failed.
[31]	7WiseUp dataset includes ninety-four students	Deep Neural Network (DNN), LSTM, DenseLSTM, RNN, Learning Management System (LMS) System	- RNN accuracy: 85%. - DenseLSTM accuracy: 96% - DNN accuracy: 67%. - LSTM accuracy: 93%.