1. The document describes a fast unsupervised approach for image segmentation based on clustering algorithm fusion and level set methods. 2. It first extracts features from images using color transformation to CIE L*a*b* color space and non-linear diffusion. Then it applies multiple clustering algorithms including Fuzzy C-means, K-means, self-organizing maps, and Gaussian mixture models. 3. The results from the clustering algorithms are fused to generate a cluster map which is then used to evolve the level set contour for segmentation. Simulation results on texture images show the proposed method achieves better segmentation accuracy compared to traditional level set approaches.