Feature Dimension Reduction for Content-Based Image Identification by Rik Das Sourav De Siddhartha Bhattacharyya

Author:Rik Das,Sourav De,Siddhartha Bhattacharyya
Language: eng
Format: epub
Publisher: IGI Global

The initial segmentation may simply be all pixels, i.e., each pixel is a segment by itself. The heart of the merging approach is the similarity criterion used to decide whether or not two segments should be merged. This criterion may be based on grey value similarity (such as the difference in average grey value, or the maximum or minimum grey value difference between segments), the edge strength of the boundary between the segments, the texture of the segments, or one of many other possibilities. The basic form of image segmentation using splitting is:

• Obtain an initial segmentation of the image

• Split each segment that is inhomogeneous in some respect (i.e., each segment that is unlikely to really be a single segment).

• Repeat step 2 until all segments are homogeneous.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.