Research on Image Segmentation Based on Fuzzy Theory

Image segmentation is one of the key steps in image analysis, where the fuzzy theory based methods are widely used, but, none had universally property to segment all types color images. So, to improve segmentation quality and universally property of segmentation algorithm, a new fuzzy color image segmentation algorithm was brought out based on feature divergence and fuzzy dissimilarity. The algorithm measured the otherness of two stylebooks space eigenvector by feature divergence, and extracted sub-images feature eigenvector using watershed algorithm, depressed operation data number. Combination of sub-image was done by dint of fuzzy dissimilarity and morphological theory, and the isolated points were eliminated, and made color image segmentation more according with the human segmentation strategy. The method efficiency and feasibility were confirmed by experimental results.