Information Extraction of Optical Answer Sheet (LJK) Based on Image Processing Using Smartphone Camera

The use of Optical Answer Sheets (UK) has become increasingly common in various educational institutions, non-governmental organizations, governmental organizations, and private enterprises. 2B pencils are typically used to indicate the answer selected in an LJK. There are many methods to scan an LJK that usually consist of a computer program (software) and a scanner. However, the increasingly widespread use of laptops and smartphones with a built-in camera can offer new possibilities for scanning without the use of a scanner. This paper presents a simple and innovative method to scan an LJK, i.e., using a built-in camera in smartphones to take an LJK image and a laptop to run the information extraction algorithm. First, the LJK image captured was extracted to display the Red component of RGB color space. Next, the Red component quality was improved using contrast stretching before converted into a binary image. The Mathematical Morphology operation of opening and closing was then applied to remove artifacts. Last, a simple grading algorithm based on geometrical position was applied to compare answer key and submitted answer from participant. The proposed method was tested using 60 LJK image samples and the results indicated that the scanning accuracy was 100% and the average computational time was 1.42 seconds, thus 11.5 times faster than the results of the previous study. Furthermore, the proposed method was also able to extract information from an LJK although test takers did not fill the LJK using a 2B pencil.