DEVELOPMENT OF BIOMETRIC EXAM FRAUD DETECTION SYSTEM USING ARTIFICIAL INTELLIGENCE TECHNIQUE

Subject Area: BIOMETRIC TECHNOLOGY

This paper presents the development of biometric exam fraud detection system using artificial intelligence technique. This aim is to employ facial recognition to detect exam fraud in tertiary institutions. This was achieved using research methods such as data collection, computer vision, face detection, data processing, Machine Learning (ML) and facial recognition. The system was designed using mathematical equations, pseudopodia and universal modeling diagram which presented the modeling of the face detection system. The ML algorithm used is the K-Nearest Neighbor (K-NN) and was trained with clusters of the data collected to generate the facial recognition algorithm used to build the exam fraud detection system. The performance of the K-NN when evaluated showed classification area under curve result of 0.83. Fraud detection system was tested with 30 self volunteered student faces and 24 were correctly classified leaving an accuracy of 80%.