Unsupervised Fuzzy C-Mean classification of texture images using improved texture energy measure

Texture is a very important attribute in image analysis. An important task of the image analysis is to segment the given image into more meaningful regions and to label the individual regions. Texture is characterized by the spatial distribution of intensity levels in a neighbourhood. This paper proposed "improved texture energy measures" for feature extraction, the images are convolved using the Laws masks. The symmetric features are combined and normalized; the features are classified using Fuzzy C-MeaTexture is a very important attribute in image analysis. An important task of the image analysis is to segment the given image into more meaningful regions and to label the individual regions. Texture is characterized by the spatial distribution of intensity levels in a neighbourhood. This paper proposed "improved texture energy measures" for feature extraction, the images are convolved using the Laws masks. The symmetric features are combined and normalized; the features are classified using Fuzzy C-Means. The results for normalization and non-normalization of feature classifications are analyzed. The experimental evaluation of the features is reported using texture database from brodatz album and the results are promising.ns. The results for normalization and non-normalization of feature classifications are analyzed. The experimental evaluation of the features is reported using texture database from brodatz album and the results are promising.