It is natural way to write Chinese characters by digital pen for foreign students, whose handwriting information is much richer than digital image. Stroke matching is the prerequisite to analyze handwriting errors of Chinese character. Present research hardly delivers the optimal solution of the problem on the growing sizes and complexity because of wide differences among learners' writing qualities and features. This paper proposes an approach based on genetic algorithm to match strokes. Construction of fitness function considers structural and writing features of Chinese characters. The method can achieve correct matching stroke rate 90.17% at least in the experiments, which indicates that our proposed approach Is effective In next steps of handwriting errors analysis.

http://ieeexplore.ieee.org/iel7/7879539/7888206/07888284.pdf

A method of matching strokes based on genetic algorithm

On-line Signature Verification is a field of verifying the time series signature data that normally obtained from the tablet-based device. Unlike common signature image, the on-line signature image data consists of points that are arranged in sequence based on time. The aim of this research is to develop a new approach to map the strokes in both test and reference signatures as well as to verify the originality of the test signatures. Current methods make use of the DTW algorithm and its variant to segment them before comparing each of its data dimension. This project suggesting a modified DTW algorithm with the proposed Missed Nodes Recovery Algorithm aims to improve the mapping performance, hence the development of stroke to stroke signature comparison is possible. This project is also proposing a method to compare the strokes with its similar strokes in on-line signature. All algorithm and experiments will be carried out using Matlab. The output of this research project is an algorithm that can be used to map strokes as well as to compare similarity of strokes in on-line signature.

Azhar Ahmad Jaini

Papers

Stroke-to-stroke matching in on-line signature verification