Pictorial pattern recognition

Some of the techniques which can be used to automatically classify and describe pictorial patterns are reviewed, with emphasis on methods which may be useful in processing biomedical pictures.

Abstract

This paper reviews some of the techniques which can be used to automatically classify and describe pictorial patterns. It discusses representative “preprocessing” and “property measurement” methods, with emphasis on methods which may be useful in processing biomedical pictures. (The decision-making aspect of pattern recognition is not discussed.) Techniques for subdividing pictures into regions, and measuring geometrical properties of such regions, are also considered. The paper is intended to serve as an introduction to the extensive literature in this field; it includes numerous references to this literature.