TEXTURE CLASSIFICATION USING A FUZZY TEXTURE SPECTRUM AND NEURAL NETWORKS

In the research areas in computer vision, many applications have been discovered using texture classification techniques, such as the conten t retrieval in multimedia, the computer-aided diagnosis of medical ima ges, and the segmentation of remote sensing images. The success of the texture classification of a given set of images hinges on the designs of texture features and the classifiers. We present a new texture fea ture, fuzzy texture spectrum, for texture classification, which is bas ed on the relative gray levels between pixels. A vector of fuzzy value s is used to indicate the relationship of the gray levels between the neighboring pixels. The fuzzy texture spectrum can be considered as th e distribution of the fuzzified differences between the neighboring pi xels. It is an improved version of the reduced texture spectrum, and i t is less sensitive to the noise and the changing of the background br ightness in texture images. We use 12 Brodatz texture images in the si mulations to show the effectiveness of the new texture feature. Our si mulation results show that the rate of classification error can be red uced to 0.2083%. (C) 1998 SPIE and IS&T. [S1017-9909(98)00301-8].