Self-adaptive image segmentation method based on quick global K-means

Abstract

The invention discloses a self-adaptive image segmentation method based on a quick global K-means, mainly solving the problems of poor effect and high calculation complexity of self-adaptively segmented images in the prior art. The self-adaptive image segmentation method is implemented through the following steps of: (1) reading an image to be segmented, and extracting texture characteristics of the image to be segmented; (2) setting a search range of an image cluster number c; (3) clustering the texture characteristics into c classes by using an improved quick global K-means method so as to obtain a clustering center; (4) calculating the membership grade of each texture characteristic belonging to each class; (5) calculating effectiveness indexes of L(c), L(c-1) and L(c-2) respectively corresponding to cluster number of c, c-1 and c-2 according to the membership grades and the clustering center; and (6) outputting a segmentation result if L(c-1) is more than L(c-2) and L(c-1) is more than L(c), otherwise, assuming c=c+1, and returning to the step (3). Compared with the other methods, the self-adaptive image segmentation method disclosed by the invention has the advantages of low calculation complexity, more accuracy of self-adaptively obtained optimal class number and segmentation result and capability of being used for segmenting and clustering images.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an image segmentation method, which can be used for image segmentation and clustering. Background technique [0002] Image segmentation is the technology and process of dividing an image into several specific regions with unique properties and proposing objects of interest. It is a key step from image processing to image analysis and is a basic computer vision technology. [0003] The problem of image segmentation can also be regarded as the problem of object classification, so the pattern classification technology in pattern recognition can be used. Image segmentation using feature space clustering method is to represent the pixels in the image space with corresponding feature space points, segment the feature space according to their aggregation in the feature space, and then map them back to the original image space to obtain the segmentation result. Among them, K-me...

AI Technical Summary

Problems solved by technology

However, these techniques have not been widely accepted

In 2003, Likas et al. proposed a global K-means clustering algorithm, but the algorithm has a large amount of calculation and poor time performance. Therefore, in his article, he also gave a fast global K-means clustering algorithm, but the algorithm There are still problems of high space complexity and high time complexity, and it is not suitable for datasets with large amounts of data

However, their evaluation effect on some complex cluster structures is not ideal, so it is difficult to get the correct optimal number of clusters

Images