Texture Segmentation based on Image Fidelity Indexes (Segmentação de Texturas Baseada em Índices de Fidelidade de Imagem)

Vinícius Jucá (vmj@dsc.upe.br)1, Carlos Mello (cabm@ieee.org)2

1Universidade de Pernambuco
2University of Pernambuco

This paper appears in: Revista IEEE América Latina

Publication Date: June 2011
Volume: 9,   Issue: 3 
ISSN: 1548-0992

Texture segmentation and analysis is the first step in the modelling if human vision to machines. Using texture descriptors, as Gray Level Co-occurrence Matrix, Local Binary Pattern and Sum and Difference Histograms, this work applies fidelity indexes to segment the textures in an image. This method is evaluated using a database with 1,452 images divided into 6 classes of different textures (242 images per class) and it was also used to improve the segmentation achieved ny quadtree decomposition. The results were always better than the classical version of the algorithm. In another experiment, the method was applied in the segmentation of 18 texture mosaics created using Brodatz textures. In every experiment, the results were very satisfactory, confirming the robustness of the proposal.

Index Terms:
texture analysis, image fidelity, image segmentation   

Documents that cite this document
This function is not implemented yet.

[PDF Full-Text (523)]