Super-pixel spectral clustering color image segmentation method based on semi-supervision

Abstract

The invention discloses a super-pixel spectral clustering color image segmentation method based on semi-supervision. The method comprises the following steps: step 1, inputting a to-be-processed RGB color image; Step 2, initializing a super-pixel initialization size and a standard coefficient of the SLIC parameter, calculating a super-pixel, and extracting a mean value of pixel points of each super-pixel block to generate a super-pixel image; 3, acquiring semi-supervised information in the super-pixel image in a manual marking and scribing mode; Step 4, constructing similarity among the super-pixel blocks by using the semi-supervised information obtained in the previous step; 5, performing color image segmentation by using an NJW spectral clustering algorithm in combination with the constructed semi-supervised similarity matrix; And step 6, obtaining a final segmentation result of the input image according to a division result of the super-pixel region. According to the method, the image segmentation speed can be increased, the algorithm efficiency is improved, and an ideal segmentation result is obtained.

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a color image segmentation method based on semi-supervised superpixel spectrum clustering. Background technique [0002] With the development of science and technology, processing the received color images for further analysis and use has become an urgent task in the development of image engineering. Image segmentation is the basis of image processing, and its results directly affect the accuracy of image analysis. The definition of image segmentation is the process of dividing the input image into several regions with special significance and distinguishing the image target and background. The traditional image segmentation technology is based on the pixel level, which often makes the final image segmentation result prone to fragmentation. In order to highlight the regional information of the image, the concept of super pixel is introduced. [0003] The image segmen...

AI Technical Summary

Problems solved by technology

However, the traditional spectral clustering algorithm is based on graph theory, and there are still many unsolved problems in massive data calculation and similarity construction. Therefore, the research on spectral clustering algorithm is still a hot and difficult point.

Images