Fuzzy clustering image segmentation method based on local information and non-local information of pixels

Abstract

The invention discloses a fuzzy clustering image segmentation method based on local information and non-local information of pixels and belongs to the technical field of image processing. The method comprises steps of extracting gray scale characteristics and neighborhood characteristics of pixels in given images so as to obtain characteristic information sets of the images; randomly generating membership degrees of the pixels; designing clustering center pairs of the segmentation algorithm and designing segmented energy functions; through an iteration process, maximizing the segmented functions and in the iteration process, based on the lagrangian function method, updating the membership degrees of the pixels and the clustering center pairs; and finishing the iteration process and based on the biggest membership degree principle, carrying out defuzzification on themembership degrees of the pixels, thereby outputting final segmentation results for segmentation of the given images. According to the invention, the neighborhood information of the pixels can be effectively used; details of the image segmentation are kept; non-local information of the pixels in the images can be sufficiently used; and robustness of the algorithm is improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a fuzzy clustering image segmentation method based on pixel local information and non-local information. Background technique [0002] Image is an important source for human beings to obtain information from the objective world and an important medium for transmitting information. With the development of technologies such as computers and the Internet, digital image processing technology is playing an increasingly important role in industries such as industry, medical care, military affairs, and transportation. [0003] In order to effectively use the relevant information in the image, it is necessary to segment the image and extract the region of interest in the image. In essence, image segmentation is to divide the image into disjoint parts, and the pixels in each part have the same or similar features, where the features can be color, texture, etc. Image segmentation...

AI Technical Summary

Problems solved by technology

[0007] The traditional FCM algorithm does not use the neighborhood information of the pixels when segmenting, and the segmentation effect is not ideal when segmenting the noise containing the image, and the noise is ubiquitous in the image, so the design has a strong robustness. Segmentation algorithm is the research focus and difficulty of image segmentation

Images