An Adaptive Approach for Extracting Texture Information and Segmentation

Texture plays an important role in visual information processing since texture provides useful information about shape, orientation, and depth of the objects. The human visual system utilizes texture without difficulty as one of the visual cues for image interpretation, scene analysis and object recognition. However, to extract and to analyze texture are difficult tasks in machine perception including computer vision and digital photogrammetry. To develop a texture analysis system capable of dealing with all aspects of texture is a very difficult visual information processing task, because of the complex nature of texture. It is the lack of mathematical models that make automatic description and recognition of texture patterns a complex and, as yet, unsolved problem. It is important to utilize properties of texture and to understand how the human visual system works for texture discrimination and grouping. In this study, an adaptive strategy is addressed by using two-dimensional Gabor functions in order to extract texture information and to segment images. The Gabor filters are conceived as hypothetical structures of the retinal receptive fields. Therefore, to develop a texture analysis system which resembles the performance of human visual perception is possible using the Gabor filters. A scheme to select appropriate texture parameters without visual inspection is discussed.