Clothing segmentation using foreground and background estimation based on the constrained Delaunay triangulation

Computer vision is widely used in many computer task , such as vision computer, such as dreseed people detection, identification, image editing, human sketches and portrait for graphics rendering. First step in image segmentation for final project is find edge from image use edge detection algorithm. And then to divide foreground and background use Constrained Delaunay Triangulation (CDT) method. With use CDT we can find clothing region that have been segmented with good accuration. From result of experiment using this CDT method obtained accuration for most optimal threshold selection, that is 83%. This is caused by difference of background and lighting variation. Whereas for result of experiment the most optimal value of hue selection is 0-0.1 with percentage attain unto 100%.