Image Segmentation Using Local Classification Of Surfaces On The Basis Of Normal Vectors Arrangement In Point's Vicinity

摘要

This paper presents a novel approach to characterization of surface shape in local area. In contrast to well-known curvature settles features proposed classification is based on the analysis of arrangement of normal vectors in the point vicinity. It is shown that the sum of corners forming normal vectors and the one from image plane could characterize the shape of surface.