An Improved Image Thinning Algorithm and Its Application in Chinese Character Image Refining
Wang Miaomiao, Zhenglin Li, Fuyuan Si, Guan Lei · 2019 IEEE 3rd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · 2019
Chinese character image thinning helps to reduce redundant information and maintain the main structure information of the original Chinese character. It has important applications in the field of character recognition. However, the current image thinning algorithm will cause new distortion in thinning, thus increasing the interference and difficulty in recognition. And there are still some problems such as slow thinning speed. In view of the above problems, this paper proposes a fast thinning method for Chinese characters. Based on the pixel-by-pixel thinning (OPTA) algorithm, the algorithm reconstructs 14 neighborhood elimination templates and retention templates, and solves the distortion and incompleteness of thinning. In addition, the algorithm also improves the recognition order to improve the speed, it give priority to the right adjacent pixels of the target point P and the down adjacent pixels. Experiments on Chinese character images show that the algorithm can refine the ridges completely and the speed is faster than other algorithms.