In this paper we propose a novel approach for image segmentation using adaptive tree-structured wavelet transform for texture analysis. We first split the input image into NxN blocks, calculate the distances between neighbor blocks by the energy signatures of the wavelet coefficients of the adaptive tree-structured wavelet transform of each block. Then we merge blocks with smallest distances to form larger regions. The process will repeat until we got desired number of regions. Experimental results show that our proposed method outperforms the existing image segmentation method. © 2009 IEEE.