Research on Segmentation Technology of Fuzzy Image with Adhesion Pixels

ABSTRACT

In a blurred image, the difference of gray level between adhesion pixels and the pixels of background image is small, so it is difficult to accurately segment. Traditional methods for segmenting fuzzy image with adhesion pixels assume that the image is large, and then the computational complexity dramatically expands, which reduces the accuracy of segmentation. In order to solve this problem, this article puts forward a method of fuzzy image segmentation with adhesion pixels based on similarity probability. Using the characteristic values of the Hessian matrix to detect the adhesion pixels, the probability whether the pixel in the image belongs to adhesion pixel can obtain. Applying the method of Gaussian gradient convolution processing, the edge enhanced image of adhesion pixel area can be achieved. Taking advantage of circular probe, the accurate adhesion points can be obtained, so as to complete accurate segmentation of adhesion pixels in the blurred image. The experimental results show that the algorithm presented in this paper for segmenting fuzzy image with adhesion pixels can greatly improve the segmentation accuracy.

Keywords: