An Improved Cost-Efficient Thinning Algorithm for Digital Image
Liang Jia, Zhenjie Hou · 2013
Digital image processing plays an important role in our everevolving ubiquitous information society with an increasing need of automatic and efficient information collection. Thinning algorithms have been widely developed and applied in Optical Character Recognition (OCR) to eliminate the redundant data as well as keep the essential features of digital images. Inspired by the in-depth analysis of results obtained from Davies’s classical algorithm, this paper proposes an improved and cost-effective thinning algorithm to enhance the accuracy of digital image skeletonization and also maintain the computation complexity at a low level. The extensive experiments show that the results of this improved thinning algorithm is able to achieve the similar accuracy of other advanced algorithms and inherits the advantage of low complexity of Davies’s classical algorithm.