摘要
针对字符图像中笔画内部像素点灰度的屋脊型分布特点,提出了一种新的字符图像细化方法。该方法利用字符图像中灰度截面曲线,获取了屋脊型边缘的候选点集,通过对候选点集的进一步筛选和修补,取得了定位较为准确的屋脊边缘点,实现了数字字符灰度图像的可靠细化。
A new method for character thinning is presented based on ridge features of character strokes. By analyzing the curve of gray values at each cross section in character area, the set of possible ridge edge points is obtained. Then the criteria for selecting ridge points with high location accuracy from the candidate set are discused. This method has been proved reliable and effective by experiment results.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2000年第3期332-337,共6页
Pattern Recognition and Artificial Intelligence
