Optimizing the correction of MCQ test answer sheets using digital image processing

Abstract

Today, multiple choice tests due to their benefits have become one of the most common methods to evaluate student's learning. In order to correct their answer sheets by the machines, very expensive equipment is necessary. Despite high correction accuracy of this method, because of some limitations, it's not suitable for public use such as schools. To overcome these limitations, attempts have been made to provide alternative correction system, but the main challenge in the machinery correction method and proposed alternative correction systems is that if the tester marks the option on the answer sheet incompletely, it will not be corrected properly. This recognition error affects on the correction accuracy. This article proposes an optimized correction method for correcting multiple choice test answer sheets using mathematical morphology, thresholding and neighborhood that with the increase of the recognition rate for incompletely marked options will improve overall correction accuracy.