Efficient rotation invariant Gabor descriptors for texture classification

In texture classification experiments, the conventional Gabor representation of textures and its extracted descriptors often yield a poor performance in classifying textures at rotated viewpoints. This paper presents a theoretically very simple, yet efficient approach for generating rotation invaria...

Full description

Saved in:
Bibliographic Details
Main Authors: Rahman, M. H., Pickering, M., Kundu, D.
Format: Conference Proceeding
Language:English
Subjects:
Brodatz
Convolution
DT-CWT
Gabor filters
Manganese
rotation invariance
Smoothing methods
sorted distribution
texture classification
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In texture classification experiments, the conventional Gabor representation of textures and its extracted descriptors often yield a poor performance in classifying textures at rotated viewpoints. This paper presents a theoretically very simple, yet efficient approach for generating rotation invariant descriptor representation by sorted distribution of coefficients (SDC) of the Gabor filter outputs smoothed by a Gaussian windowing function. The classification performance is tested on a set of 112 textures from Brodatz album where each texture is rotated in 7 directions. Our implementation exceeds the best reported results and achieves comparable performance on the rest. Our experiments demonstrate that the image representation based on SDC is more effective in classifying textures rotated at different angles.
DOI:10.1109/ICIEV.2012.6317469