Automated Verification Method of Korean Word Handwriting Using Geometric Feature

Handwriting is a behavioral characteristic of an individual, and is used as a biometric method for identity verification. The handwriting comparison can not be trusted because the handwriting similarity is judged by the subjective criteria of the expert. In this paper, to compare two images written with the same Korean word, some digital image processing techniques are used to detect geometric feature points such as corner, ending, and bifurcation points of a stroke. Then, adjacent redundant feature points is removed. Word images of different sizes are normalized by multiplying the coordinate values ​​of the feature points by the magnification of the image. The similarity between two sets of feature points is measured through modified Hausdorff distance. In the case of genuine matching, it is confirmed that there is a correlation between the number of feature points and distnace. Therefore, the calculated distance is normalized by dividing the number of feature points by the number of feature points. Experimental results show the possibility of automated handwriting verification by showing 37% EER.