Hierarchical Texture Segmentation Using Dictionaries.

P. Bajcsy N. and Ahuja

3rd Asian Conference on Computer Vision, Hong Kong, January 8-11, 1998.

We present a new hierarchical texture segmentation method that partitions an image into textured regions. A textured region is viewed as a set of uniformly distributed primitives. A primitive is a region with constant grey values. Gray values within a primitive can be corrupted by noise. Any noisy primitive contains grey values from a δ-wide interval (δ-homogeneous primitive). The noisy primitive is described by the mean of interior gray values. A textured region with the noise is characterized by a set of gray value means (texture dictionary) derived from noisy primitives.

Every pixel (sample point) and its neighborhood give rise to an estimate of texture dictionary. Components of the estimated dictionary at a pixel characterize noisy primitives of a textured region grown from the pixel.

Co-occurance of noisy primitives from this grown region are calculated. Final segmentation is obtained by grouping pixels with identical dictionaries and co-occurances created at each pixel. Homogeneity degree δ of noisy primitives provides a framework for multiscale analysis. Computational efficiency and robustness of the proposed method are related. Experiments are reported for synthetic and real textures from Brodatz album and real scenes.