Computer processing of line images: A survey

摘要

This paper investigates the state of the art in computer analysis of digitized line images, with emphasis on the requirements of character recognition. The main discussion concerns the requirement of converting a bit mapped, grey-level or binary image to a polygonal approximation of the “skeleton” or “medial axis” of the character.A wide range of thinning algorithms are outlined, followed by some methods of polygonal approximation. The paper concludes that without cheap highly parallel processing power, the common iterative methods are of no use in a cost effective OCR system.