Document image segmentation using recursive top-down approach and region type identification

Abstract

Document image analysis refers to algorithms and techniques that are applied to images of documents to obtain a computer-readable description from pixel data. This paper presents a Top-down approach for document image segmentation based on recursively finding rectangular blocks in the image. The proposed algorithm recursively finds the rectangular regions in the image document using vertical and horizontal profiles and each identified block is further analyzed to identify its type whether it is text, picture or table. The method used is not language specific. Documents of different languages have been tested and satisfactory results have been obtained. In this paper we have also briefly described the existing algorithms and the methods that are used for document segmentation. © 2010 Springer-Verlag Berlin Heidelberg.