期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2014
卷号：63
期号：2
出版社：Journal of Theoretical and Applied
摘要：Thinning or skeletonization is become a curial step in many document image analysis and recognition application as a pre-process stage. These applications such as optical character recognition OCR, optical script recognition OSR, and optical font recognition OFR widely adopt many exist methods, some thinning challenging defect the performance of these applications. However, further enhancement for thinning technique is indeed. A simple iterative thinning algorithm is proposed for binary images domain. Thereafter, a benchmark dataset is designed for thinning algorithm evaluation is used. Finally, The evaluation experiments result are compared with some other major thinning methods to proof the proposed method efficiency. The proposed algorithm visual experiments shows outperform for many thinning challenges such as one pixel width, topology preserving, rotation tolerance, and tail occurrences.
