摘要
一个优异的算法是OAS中图像识别的关键,Zhang快速并行细化算法具有快速实用的特点,同时能保证细化后曲线的连通性,但细化后的骨架不能保证为单像素宽,这给找细化图像的分叉点带来了困难。为此文章在Zhang算法的基础上进一步对细化后的骨架单像素化,同时保持原有算法的优良特性。改进的细化算法应用于手写体汉字串的细化,取得很好的细化效果。
Efficient algorithm is key to computer graphics in OAS, and Zhang＇s fast parallel thinning algorithm is recognized as an easy and efficient algorithm. To the thinned curves, it can ensure graph connection, but cannot ensure the singular pixel width, which brings difficulty for determining fork points on the thinned images. Aimed at this deficiency, we extend Zhang＇s algorithm to make thinned curves into singular pixel width on the premise of maintaining the advantageous features of Zhang＇s algorithm. Finally the extended algorithm is applied to the string thinning of Chinese handwriting characters with good effect.
出处
《福建工程学院学报》
CAS
2006年第1期89-92,共4页
Journal of Fujian University of Technology
基金
2004年福建省教育厅B类科研项目(JB04161)
