Morphological thinning based on image's edges

Thinning algorithms are widely used in image processing and pattern recognition. The skeleton obtained from them not only maintains the geometric and topologic features of the original image, but substantially reduces its redundancy also. This paper analyzes some characteristics of the skeleton obtained by morphological erosion and opening. An improved thinning method-morphological thinning based on image's edge is presented. The global skeleton got from it has much less discontinuities. Importantly, it's just one pixel wide. Finally, from skeleton subsets, the original image can be precisely reconstructed.