Economical Solution to Automatic Evaluation of an OMR Sheet Using Image Processing

Abstract

Optical Mark Recognition (OMR) Sheet, also called as Bubble sheet is a special type of form used to answer graded multiple choice question examinations, where students have to mark or darken the bubbles to answer the questions. OMR sheets are used by various school, college, university and competitive examinations. Larger institutions like universities have specialized machines to optically detect the marked answers and grade the student. However, these scanners are significantly heavy on the pocket and cannot be afforded by small tuitions or individual teachers. With the help of mobile phones, it is easily possible to scan multiple sheets within seconds. The user is expected to upload a single file of all responses and view the results in a tabular format. Using various image processing techniques, the algorithm evaluates the responses of the student and displays the grade and percentage on the input image along with the correctly, wrongly marked and actual answers with green, red and yellow colors respectively. The user also has access to an interactive dashboard which presents various analytics regarding the students’ performance. A unique feature in this implementation, is that users can choose the number of questions, number of choices, marks per question etc. after which he/she would be able to download a customized OMR response sheet. We experimented on 300 sample OMR sheets with 12,800 questions and obtained an average accuracy of 99.12%. This paper aims to present a cost-effective solution to accurately scan OMR sheets without the need of scanners.