Linear Regression Model on Multiresolution Analysis for Texture Classification

Abstract

Texture is a surface property which is used to identify and recognize the object. Texture analysis is important in many applications of computer image analysis for classification and segmentation of images based on local spatial patterns of intensity or color. In texture classification the goal is to assign an unknown sample image to one set of known texture classes. The proposed method is texture analysis and classification with linear regression model based on directional lifting based wavelet transform. In this method, texture classification is performed by analyzing the spatial correlation between some sample texture images belonging to the same kind of texture at different frequency regions, obtained by 2-D wavelet transform. The linear regression model is employed to analyze this correlation and extract texture features that characterize the samples. Therefore, this method not only considers the frequency regions but also the correlation between the frequency regions. So the classification rate is improved.