Exam Answer Detection using Proposed Nested Loop Pixel Checking Based on Image Processing

Abstract

Tracking object is one of the most discussed problems in digital images processing. The objectives of tracking objects include the counting, correction, classification and so on. This study discusses how tracking objects, in this case the multiple choice chosen by students, is carried out on the exam answer sheet using the proposed nested loop pixel checking method. The answer sheet used here is the answer sheet for the exam from SMK Khoiriyah Hasyim Tebuireng Jombang because each school has a different exam answer sheet format so that the coordinate parameters of each answer sheet between schools must be different. Before the nested loop pixel checking method, preprocessing was carried out first after taking the image, namely cropping the Region of Interest, conversion to grayscale, filtering, and edge detection using the adaptive Canny edge detection method. The process of nested loop pixel checking is done by iterating successively with steps per answer option, per number of answers, and per column. In addition, a special case is given to find out whether this proposed method works as desired. The results show that this method can track answers for both normal and special cases.