Forma, Vol. 1 (No. 1), p. 84, 1985

Shape Features in Pictorial Pattern Recognition

Jun-ichiro Toriwaki

Department of information Engineering, Faculty of Engineering, Nagoya University

Abstract. In this paper, typical shape features used in pictorial pattern recognition is briefly surveyed. First basic viewpoints of shape feature analysis in the context of pattern recognition is presented with comparison to those in the science of form in general. Then, after short reference to digitization of pictures, author briefly introduces major shape features which have been developed in the field of imagery pattern recognition. Following items are referred to in this article: digital geometry, thinning, distance transformation and skeleton, border-following, descriptor of positional relation, features of spatial distribution of many figures, contours and structural lines (ridges, ravines, top, etc.) of a grey pictures. and geometrical features of a single figure.