Mean shift and fuzzy clustering-based natural image unsupervised segmentation method

Abstract

The invention provides a mean shift and fuzzy clustering-based natural image unsupervised segmentation method, and mainly solves the problem of low unsupervised segmentation accuracy of massive natural images in the prior art. According to the scheme, the method comprises the steps of 1) inputting an image and smoothing the image; 2) uniformly initializing 64 iterative initial points in a normalized RGB color space of pixels of the smoothed image; 3) performing an iterative search on the initial points to obtain 64 convergence points; 4) deleting the convergence points with pixel numbers smaller than a deletion threshold in high-dimensional balls taking the convergence points as centers; 5) combining the convergence points with Euclidean distances smaller than a combination threshold, determining density peak values and a density peak value number, and calculating membership degrees of the pixels and smooth membership degrees of the pixels in sequence; and 6) defuzzifying the smooth membership degrees of the pixels, adding class tags to the pixels, and outputting segmented images. According to the method, control parameters do not need to be set; a segmentation type number of the image can be automatically determined; and the method can be used for unsupervised segmentation of the massive natural images.

Description

technical field [0001] The technical field of image processing of the present invention is specifically an unsupervised segmentation method for natural images, which can be used in video object tracking and recognition and content-based image retrieval. Background technique [0002] In recent years, with the rapid development of science and technology and computer Internet technology, digital images are used more and more widely in all walks of life. How to quickly identify an image in a large number of images has always been a hot topic of discussion in computer vision and pattern recognition. Since the content-based image retrieval technology was proposed in the early 1990s, it has been a research hotspot for researchers. It mainly extracts the texture, color, shape of the target and their spatial position information of the image, and calculates the information obtained by the image. Retrieve the similarity distance between the image and the image in the data set to real...

AI Technical Summary

Problems solved by technology

For natural images with simple structures, better methods can accurately determine the number of categories, but for images with more complex structures, there will generally be errors

The existing unsupervised segmentation methods of natural images are unstable, which can easily cause over-segmentation or under-segmentation of images

Due to the complexity of natural image scenes, there is no better method to accurately determine the number of categories of natural images

Images