Extraction and Merging of Stroke Structure of Chinese Characters

Chinese characters are complex graphics with strokes as the basic unit. In order to analyze their structure, stroke extraction is the first step. This paper presents an automatic extraction method of Chinese character strokes, which regards the extraction of Chinese character strokes as finding the optimal path and merging. Based on the superpixel network, the path network is applied to enumerate all possible stroke segments. Then, the repeated strokes are merged according to the Intersection over Union. Experimental results show that the method can effectively extract accurate strokes and obtain single strokes with high-level semantic features and complete information.