Figure-Ground Separation for Vector Graphics with Contour Detection

T. Hayashi, K. Shibata, R. Onai, and K. Abe (Japan)

Keywords

vector graphics, figureground separation, image segmen tation, contour, active contour model

Abstract

This paper reports a figure-ground separation method for vector graphics. Although figure-ground separation is an effective model to mirror human’s perceptions to image segmentation, no figure-ground separation method for vec tor graphics has been studied. In this paper, we propose a novel figure-ground separation method for vector graph ics. The method conducts figure-ground separation without rasterizing vector graphics. First the method finds the prim itives that express the boundary between the figure and the ground in a vector graphic. Next the method extracts inner regions of the boundary as the figure and the other regions as the ground. We have evaluated the effectiveness of the method by comparing it with a traditional method for raster images using 126 vector graphics. The experimental results have shown that 81.7% of the vector graphics are correctly separated figure and ground by the proposed method and that this performance is 37.2% higher than the traditional method.

Important Links:



Go Back