Fuzzy feature extraction on handwritten Chinese characters
Hahn-Ming Lee, Chiung-Wei Huang · 1994
In this paper, a new feature extraction method based on the fuzzy set theory is proposed. Our feature extraction focuses on decomposing the structure of handwritten characters. Due to the Chinese characters are composed of strokes, features adopted here are all strokes. During extraction, extracted strokes are assigned various degrees for each stroke type. Also, some flexible fuzzy rules are given to deal with the strokes combination. Since the strokes of handwritten Chinese character are fuzzy in nature, our approach can obtain a better result on depressing the writing variants and distortions which are caused by thinning. Moreover, current simulation results encourage us to do further investigation.>