Local Segmentation of Touching Characters using Contour based Shape Decomposition

We propose a contour based shape decomposition approach that provides local segmentation of touching characters. The shape contour is linearized into edgelets and edgelets are merged into boundary fragments. Connection cost between boundary fragments is obtained by considering local smoothness, connection length and a stroke-level property Similar Stroke Rate. Samples of connections among boundary fragments are randomly generated and the one with the minimum global cost is selected to produce optimal segmentation of the shape. To obtain a binary segmentation using this approach, we make an iterative search for the parameters that yields two components on a shape. Experimental results on a number of synthetic shape images and the LTP dataset showed that this contour based shape decomposition technique is promising and it is effective on providing local segmentation of touching characters.