Paper Title
X-Mark Detection of Multiple Answer Test using Template Matching Technique
Abstract
In this paper a template matching technique to identify the presence of X-Mark on the answer sheet of multiplechoice
type for automated scoring purpose is proposed. The proposed method will be able to reduce the cost of widely used
solution using an expensive machine called the Optical Mark Reader (OMR). The proposed method also allows students to
make corrections to their answers by crossing the X mark with minus or plus sign. There are two main stages in the X-mark
detection procedure: (i) input preparation stage and (ii) detection stage. In input preparation stage, three sets of experiment
data are used. The first set consists of 1311 images of characters X-mark, the second one has 46 images of the correction
Xplus-mark and 167 images of the correction Xmin-mark. Normalization step, as part of input preparation stage, normalizes
each image’s size to 11 x 11 pixels, and then applies shifting algorithm to that image. In masking process two types of
template with different weight are compared. Three experiments are conducted to measure the success rate of the proposed
method. The results show that the proposed method gives 94% of accuracy for the X-mark detection and 100% for the
corrected Xplus-mark detection.
Keywords - X-mark, Optical Mark Reader, Correction, Template Matching.