Image representation by self-organizing conformal network

Conformal mappings are incorporated into the self-organization model to represent images harmonically. This network is used to partition an image into quadrilateral regions, where each region contains similar features. We then map each region to a corresponding square region to unify information rep...

Full description

Saved in:
Bibliographic Details
Published in:The Visual computer 2000-01, Vol.16 (2), p.91-105
Main Authors: Tai, Wen-Pin, Liou, Cheng-Yuan
Format: Article
Language:English
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Conformal mappings are incorporated into the self-organization model to represent images harmonically. This network is used to partition an image into quadrilateral regions, where each region contains similar features. We then map each region to a corresponding square region to unify information representation and facilitate computations. This mapping is constructed to preserve spatial information while complying with the conformal property of the network. An approximated image in each square region provides us with an effective representation of the image in both modeling and compression applications. This approach has been particularly developed for large continues images.
ISSN:0178-2789
1432-2315
DOI:10.1007/s003710050199