Design of multiple Gabor filters for texture segmentation

This paper presents a method for the design of multiple Gabor lters for segmenting multi-textured images. Although design methods for a single Gabor lter have been presented recently, the development of general multilter multi-texture design methods largely remains an open problem. Previous multilter design approaches required one lter per texture or were constrained to pairs of textures. Other approaches employed ad hoc banks of Gabor lters for texture segmentation, where the parameters of the constituent lters were restricted to xed values and were not necessarily tuned for a speci c texture-segmentation problem. The proposed method removes these restrictions on the number of lters and the number of textures. This offers the potential to improve the segmentation performance or to reduce the number of lters. Further, the development of the design method and mathematical models provide new insight into the design of multiple Gabor lters for texture segmentation. Results are presented that con rm the efcacy of our lter-design method and support underlying mathematical models.