Proc. of RIAO'88, pp.1001-1008 (1988)
Abstract
An automatic segmentation method for color document images is proposed to develop a content retrieval system for scientific and technical documents containing text, color pictures in two decision stages by local and global regions. In the first stage, local decisions for each pixel are made using chromatic values in color document images. Connected components are extracted. In the second stage, global decisions are made for connected components. Document regions are classified as characters, color pictures and diagrams by the chromatic characteristics of their component images.
←[マルチメディア]
←[中野の研究紹介]
←[中野の目次]
| mail address: | |
← お手数ですが打ち込んで下さい |
First Written Before June 16, 1998
Transplanted to KSU Before May 17, 2003
Transplanted to So-net May 3, 2005
Last Update April 12, 2007
© Yasuaki Nakano 1998-2007