Using FCM for Color Texture Segmentation Based Multirscale Image Fusion

The paper presents color texture segmentation using FCM for color texture segmentation based multi-resolution image fusion. First, a color texture images are decomposed of multi-resolution representation by wavelet transform, adaptive fusion weight value of wavelet coefficients are resolved using PCA, then fused images is formed by inverse transforming and combining all wavelet coefficients, the proposed algorithm should keep the information of global structure and significant features from the color texture images. Finally, the Fuzzy C-means clustering algorithm has been used for unsupervised segmentation. The experimental results show the proposed procedure work well in the image segmentation for color texture.