A model-driven fuzzy natural stroke extractor for off-line loosely-constrained handwritten Chinese characters

A new stroke extraction approach for handwritten Chinese characters is proposed. It uses fuzzy techniques in a “hit-all” matching strategy to reduce the noise in the final output. A number of ambiguities commonly encountered in loosely-constrained handwriting are treated. A maximum of 20 distinct natural stroke classes can be extracted from each input character, together with a close estimate of the actual count of strokes which compose the character. Our system is found to have an extraction performance which is very comparable to other existing approaches. Our system offers a number of performance tuning capabilities, such as the computation of the fuzzy scores of each extracted stroke, the adjustment of the fuzzy stroke model parameters and the potential of incorporating personal writing styles into our methodology