Fast superpixel segmentation method

Abstract

The invention discloses a fast superpixel segmentation method which is a top-down iterative segmentation method. The divisibility of the region is judged by analyzing the color consistency of the regional pixel set according to the method. The method comprises the following steps that step 1: the corresponding grayscale image I<Gray> of the original color image is acquired according to the formula(1); I<Gray>(x,y)=[I<R>(x,y)+ I<G>(x,y)+ I(x,y)]/3, wherein I<Gray>(x,y) refers to the pixel grayscale of the coordinates (x,y) on the image I<Gray>, I<R>(x,y) refers to the pixel grayscale of thecoordinates (x,y) on the image I<R>, I<G>(x,y) refers to the pixel grayscale of the coordinates (x,y) on the image I<G>, and I(x,y) refers to the pixel grayscale of the coordinates (x,y) on the image I; and I<R>, I<G> and I are the red, green and blue channel images of the original color image; and step 2: the superpixel region set S is calculated according to the grayscale image I<Gray>.

Description

technical field [0001] The invention relates to the field of digital image processing, in particular to a fast superpixel segmentation method; Background technique [0002] Superpixel segmentation technology is a kind of image preprocessing technology, which clusters the pixels in the image according to the similarity of their color, texture and other characteristics, and replaces the pixel with superpixel as the smallest processing unit, which can reduce the redundant information in the image , greatly improving the speed of many image processing algorithms, which is of great significance to the development of the field of digital image processing. [0003] The existing superpixel segmentation methods can be mainly divided into two categories: top-down graph cut method and bottom-up clustering method. The first method minimizes the energy function by iteratively segmenting the image to determine the superpixel , a typical method such as the Normalized cuts superpixel segme...

AI Technical Summary

Problems solved by technology

[0003] The existing superpixel segmentation methods can be mainly divided into two categories: top-down graph cut method and bottom-up clustering method. The first method minimizes the energy function by iteratively segmenting the image to determine the superpixel , a typical method such as the Normalized cuts superpixel segmentation method proposed by Shi et al. The energy function used in this method is designed according to the contour and texture features of the image, but this method has been proved to be an NP-hard problem. The algorithm The complexity is high, and it is not suitable for processing large-scale images; the idea of ​​the second method is to first initialize a certain number of superpixel center points, and then iteratively update the clustering results until a certain convergence condition is met, so as to determine the superpixels, such as A local iterative clustering algorithm SLIC proposed by Achanta, it first initializes multiple uniformly distributed cluster centers on the image, then classifies all pixels on the image according to the color and spatial distance of the pixels, and then continuously updates the cluster centers , re-clustering until the termination condition is met to form superpixels

[0004] In order to ensure a more accurate superpixel segmentation effect, the existing superpixel segmentation methods need to pay a large time cost, which is not suitable for image analysis problems that have high requirements for algorithm efficiency.

Images