This document summarizes a research paper that proposes a new texture analysis and classification method based on wavelet transforms and linear regression models. The method analyzes the correlation between different frequency regions obtained from a 2D wavelet packet transform of sample texture images. A linear regression model is used to extract texture features characterizing this correlation. The method is compared to other multiresolution techniques like tree-structured wavelet transforms and Gabor transforms. Experimental results on a dataset of 40 textures show the new method significantly improves texture classification rates compared to existing approaches.