Method for image segmentation to identify regions with constant foreground color
作者： Timothy W. Jacobs , Zhigang Fan
DOI:
关键词:
摘要: The present invention is a method for image segmentation to produce mixed raster content (MRC) with constant foreground layers. extracts uniform text and other color objects that carry detail information. includes four primary steps. First, the are extracted from image. Next, tested consistency features decide if they should be chosen coding MRC then clustered in space. finally segmented such each layer codes same cluster.
google.ca 参考文章(6)
Kuo Chung-Hui, Gerhard R. Thompson, Ravishankar Rao, Method for combining feature distance with spatial distance for segmentation ,(2001)
Junichi Shiomi, Hiroshi Sano, Atsushi Imamura, Region segmentation of color image ,(2002)
Shriram Revankar, Zhigang Fan, Image segmentation system ,(1998)
Hui Cheng, Zhigang Fan, Background identification based segmentation and multilayer tree representation of document images international conference on image processing. ,vol. 3, pp. 1005- 1008 ,(2002) , 10.1109/ICIP.2002.1039145
Dong-joong Kang, Ji-yeun Kim, Chang-young Kim, Seong-deok Lee, Yang-seock Seo, Image segmenting apparatus and method ,(2000)
William B. Culbertson, Mark R. Stevens, Histogram-based color consistency measure for scene reconstruction ,(2001)