A Face Recognition System Based on Deep Learning (FRDLS) to Support the Entry and Supervision Procedures on Electronic Exams

The novelty of this paper is represented in using some artificial intelligence techniques in the entry control to the electronic exams (E-exam) in addition to monitoring students and distinguish the situation they are during the E-exam. Therefore, the proposed system divides into two main parts, the first part to support Eexams to handle some of the weaknesses points such as validation from students' entry by using deep learning. The Self-Organized Maps (SOM) neural network was used to recognition on students' faces. SOM is characterized by its efficient for faces' image data management, as well as it's the closest technique to match inputted untrained faces' images with a database of trained faces' images accurately. On the other part, the Bag of Words model (BoWM) is used to discriminate the status of students during the exam process. The BoWM is based on Speeded-Up Robust Features (SURF) that building on the strengths of the leading existing detectors and descriptors by using a Hessian matrix. Then extracts a report showing the status of the student such as confusion, concentration, cheating ... etc. From the experimental results, the proposed system was verified images of students' faces with high accuracy and execution time have a significant indication. Determining the status of the student during the exam by adopting the technique of retrieving documents known as the bag of word model, which proved the accuracy of determining the status of the student arrived in some cases to 100%.