Application development for multiple-choice test check by image processing
Keywords:Multiple-choices answer sheet, Fourier Convolution, Support Vector Machine
A multiple-choice examination checking (Optical Mark Recognition: OMR) machine is widely used in academy for evaluation the effectiveness of study because it has a robustness and fast for answer checking and score processing. However, this machine is quite expensive and not available for low budget academies. Moreover, the answer sheet for OMR machine is a specific. It is difficult to change an academy's answer sheet. This paper aims to develop an application for a multiple-choice answer sheet checking. The proposed method consists of two frameworks; 1) cross marker detection and 2) answer checking. For evaluation, Sakonnakorn Rajabhat University multiple-choice answer sheet (120 items and 5 choices) is implemented throughout in this experiment. The answer sheets are marked by 65 students; 40 volunteers and 25 students in class of computer and fundamental information technology. The effectiveness of cross marker detection (9,6 00 markers) provides 97% accuracy. For answer checking provides 100% accuracy.