Automated Scoring of Multiple-Choice Test Using Template Matching Technique

Abstract

In this paper, design and implementation of automated scoring of multiple-choice test using template matching technique is reported. In the proposed method an Android based smart-phone is used to identify the presence of X-mark in the answer sheet, instead of identify which circle in the answer sheet is blackened as usually do in the widely known solution using Optical Mark Reader (OMR). Cost and practicality wise, the usage of smart-phone compared to OMR is expected to increase the accessibility to automated scoring of multiple-choice test. Furthermore, the identification of X-mark opens the possibility for students to change their answer easier by making corrected X-mark and then placing a new X-mark in another blank square compared to wipe a blackened circle and filling a new circle. Four main stages are involved in the implementation, namely image preparation stage, blank squares position determination stage, X-mark detection stage, and evaluation stage. After those stages are implemented, tests are conducted to find the best distance and angle of the smart phone to the answer sheet in taking the answer sheet's images, and to evaluate the accuracy in identifying right answers and gives valid scoring. The success rate of recogntion the answer is at least 99%.