An Enhanced Model for Detecting and Interpreting Examination Impersonators' Handwriting in Nigerian Universities using Convolutional Neural Networks

摘要

The problem of examination malpractices by students of Tertiary Institutions in Nigeria has continued to increase due to impersonators and lack of innovative strategies such as the ability to compare and interpret the impersonator’s handwriting. However, there have been several existing models to detect and interpret the handwriting of examination impersonators, yet despite the achievement of these models, there are still certain anomalies that promotes examination malpractices in Tertiary Institutions. In this work, we developed an Enhanced Model for Detecting and Interpreting Examination Impersonators’ Handwriting in Nigerian Universities using Convolutional Neural Network (CNN). The methodology used is System Development Lifecycle Methodology (SDLC) in his approach. We implemented with JAVA Programming Language and MySQL Relational Database Management System as backend. The results show that handwriting recognition using deep learning technique and Convolutional Neural Network is a very powerful tool for problem solving, especially in the area of curbing examination malpractices in Tertiary Institutions. Furthermore, the total performance point of 23.0 clearly shows that our improved system outperforms other existing systems. This work could be beneficial to the Management of Tertiary Institutions in Nigeria and to any other institutions that deal with examinations since it provides relevant information on strategies involved in tracking down the examination impersonators.