Managing processes of perceptual organization for emerging geometrical objects

Three lower level layers of the hierarchical machine perception system are described, and experimental results are provided. The system is based on agile camera and aimed at automatic recognition of simple geometric scenes, made of crafted manipulation objects. The primary process of grouping the edgelets on the image is described. Two methods of edgelet grouping are compared, namely a Hough transform based and the one based on LSR. Some modifications to both methods have been proposed. Next, a process of perceptual grouping of elementary features - short segments - is described. Three approaches to the features fusion: one based on crisp, geometrical, heuristic conditions, the second based on fuzzyfied conditions and the third one based on Dempster-Shafer theory are addressed. The results obtained with three methods are compared and presented on the example image of the real scene. Higher layer issues, namely a problem of contour closing and of separating particular objects (in the case of occlusions) are briefly discussed. Some preliminary results in that area are reported.