An automated multiple choice grader for paper-based exams

Abstract

In this paper an automated multiple choice grader for paper-based exams is implemented. The system consists of two main parts, a software program and a document feeder scanner. The exam papers are fed to the scanner which scans them one by one and send them as an input to the software. The software program recognizes the student Identification Number (ID) and the answers for each exam paper and reports the final results in an Excel sheet. The system starts by applying an aligning procedure and segmenting the scanned image in order to extract form number, student ID, and answers boxes, then a pre-processing step that handles all irregular cases of input is implemented; where in this step a best possible shape that results in the highest recognition accuracy is gained. After getting a proper separated characters and numbers, a feature extraction process is applied on each character/number to calculate its feature vector. The feature vector is then compared with templates of feature vectors for each of the answers choices and numbers with their variations, where both characters and numbers are in English language. After recognizing all the answers and all ID number digits; the system starts grading the student paper and comparing student answer with the pre-entered key answers. A recognition rate of 95.58 % is attained.