AI based Examination Assessment Mark Management System

Calculation of exam scores and management of these scores is a crucial process in every examination. But these works have always been a tedious process, including calculation of scores for each question, then the calculation of total marks for each student and then updating the marks for each student. In spite of the modern technologies and new techniques around us, these chores are still done in traditional methods to avoid errors as the tasks considered are crucial. Hence, we focus to reduce the time spent in all these tasks and the human intervention in the entire process by using modern techniques and still provide the same quality in work done. We propose a system using Image Processing based edge detection and contour detection techniques along with Deep Learning based Image classification technique to automate the entire process. We perform two tasks: (1). Automated table detection from the students' answer scripts using image processing (2). Automated mark detection and the total calculation from the students' answer scripts using deep learning techniques. We use Convolutional Neural Networks (CNN) trained and tested on handwritten digit dataset obtained with a high accuracy rate to do the task of mark detection from the answer sheet. We have done our evaluations using the answer sheet used in our college and the system works well for table recognition for the given pictures.