An Efficient Parallel Thinning Algorithm using One and Two Sub-Iterations

Prashant Aparajeya and Sudip Sanyal


Thinning algorithm, Elimination condition, Connectivity preservation, Template structures


Two approaches are proposed for thinning of binarized images which work in parallel. The first approach uses two sub-iterations in which, during odd iterations top-right pixels are deleted while in even iterations bottom-left pixels are deleted. The second approach uses a single sub-iteration in which pixels are deleted when certain conditions are satisfied. Both approaches result into a thinner binary image, and finally a thinned medial curve is obtained. A 3×3 operator is used through which templates are designed for pixel deletion, while a 4×4 operator is used to preserve some template structures in the second approach. The shape of the medial curve and connectivity is maintained in the obtained thinned image. Finally straightening and noise removal from obtained curve is done so that the property of original image is maintained. Detailed comparisons with existing algorithms show the better thinned image quality, efficiency and performance of the proposed technique.

Important Links:

Go Back