A Novel Stroke Extraction Model for Chinese Characters based on Steerable Filters

The technology of stroke extraction plays a key role in the off-line Chinese character recognition field. The accuracy and computing complexity of classifying depend on the output of stroke extraction. Based on the analysis of the existing methods, this paper proposed a novel stroke extraction model with steerable filters. The experiments indicate that this model can be used to extract strokes from both printed and handwritten Chinese characters. Meanwhile, it also can be applied to eliminate the redundant stroke connections impressively with considerable low computing complexity