Rotation invariant texture classification using covariance
Sashidhar Madiraju, Chih-Chiang Liu · 2002
Proposes a simple and powerful approach for texture classification using the eigenfeatures of local covariance measures. A texton encoder produces a texture code which is invariant to local and global textural rotations. The proposed method uses six statistical features obtained from two scales of this invariant encoder to result in indices for roughness, anisotropy, and other higher-order textural features. Classification results for synthetic and natural textures are presented. The authors also discuss the effect of window sizes used at local and global scales on the performance of the classifier.>