Nonsupervision image segmentation process based on clone selection

Abstract

The invention discloses an unsupervised image segmentation method, which belongs to the technical field of image processing, and aims to reduce the dependence on user apriori knowledge as well as ensure segmentation precision. The method comprises the following steps: (1) initializing an antibody population and setting up parameters; (2) calculating an affinity degree and operating clonal reproduction; (3) carrying out a clonal variation operation on the clonal reproduced antibody population; (4) carrying out a migration operation on the clonal variated antibody population; (5) carrying out an iterative operation on the FCM; (6) calculating the affinity degree of the antibody population and carrying out clonal selection operation; (7) repeating the steps and ending the iterative operation according to iterative conditions; and (8) dividing an image according to the antibody optimal and outputting the segmentation result. The method has the advantage that the image can be automatically segmented effectively under the circumstance that the amount of the segmentation regions is unspecified.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an unsupervised image segmentation method based on clone selection, which can be applied to automatic segmentation of digital images. Background technique [0002] 80% of the information received by humans comes from visual or image information, including images, graphics, animations, videos, text data, etc. This is the most effective and important way of obtaining and exchanging information. With the popularity of computers, people are increasingly using computers to help humans acquire and process image information. Image processing, image analysis and image understanding, the organic combination of these three levels is called image engineering. [0003] As one of the basic problems of image processing, image segmentation divides the image into a set of non-overlapping regions. These regions are either meaningful to the current task, or help to explain the relations...

AI Technical Summary

Problems solved by technology

Automatic segmentation does not require human-computer interaction, and the number of segmented regions is automatically determined according to the gray information of the image. The region segmentation based on clustering technology achieves automatic image segmentation to a certain extent, but due to the convergence of the clustering algorithm Especially for images with multiple regions, it is easy to fall into the local optimal solution. This kind of segmentation method cannot guarantee the optimal segmentation results, and the reliability and stability are poor.

Images