Image segmentation is a fundamental and important step for image analysis. Tremendous efforts have been made to develop robust and efficient segmentation techniques previously. However, segmentation for texture images remains as a challenging and unresolved problem due to its textural feature. While classical approaches may fail to give successful segmentation for texture images, human vision demonstrates its incredible ability in localizing the boundaries among various textures. Encouraged by the human visual performance, an new early vision model has been proposed in one of our previous works attempting to mimic the human visual perception. This model converts a texture image into a new representation called distance map. Since the boundaries of textures are highlighted in the distance map, segmentation of a texture image turns out to be manageable. Based on the new vision model, the boundary function is described by the spatially adaptive wavelets that can represent the sharp corners as well as smooth boundaries. New global energy functions built on local texture properties are investigated. The optimization is carried out via the full power of multiresolution analysis in wavelets. Simulated annealing is applied at the same time to find the global optimization. As a result, both the improvements of image representation and segmentation technique have advanced the state of art in this area. Simulation and empirical studies on texture images have been performed. The results confirm the enlightenment.