Node2Vec algorithm-based ultra-pixel image edge detection method

Abstract

The invention relates to a Node2Vec algorithm-based ultra-pixel image edge detection method. The method includes the following steps that: 1) on the basis of a superpixel simple linear iterative segmentation algorithm (SC++), a k-means++ clustering algorithm is utilized to automatically find out an initial cluster center, and pixel points are clustered to cluster centers closest to the Euclidean distances of the pixel points, and the clustered pixel points are marked with corresponding class labels, and superpixels are extracted; 2) the gradient values, gradient directions, Euclidean distances, circumscribed circle radiuses, inscribed circle radiuses of center points are calculated, and a full connection dense network is established on the basis of the center nodes of the superpixels and similarity relationships between the center nodes; and 3) a double-threshold method and a soft threshold method are utilized to make the full connection dense network sparse; and 4) network nodes are transformed into vectors, and edge points are found out. The method of the invention has high detection efficiency; the SC++ algorithm is adopted to extract the superpixels; the double-threshold methodand the soft threshold method are adopted to make the full connection network sparse; and a Node2Vec algorithm is adopted, so that image edge detection can be realized.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an image edge detection method. Background technique [0002] Images contain a wealth of information. The shape, color, structure, edge and other information contained in it are intuitive to express, easy to disseminate and utilize information, and play a very important role in today's multimedia image processing technology. With the continuous development of computer technology, obtaining and processing image information by computer becomes faster and more convenient. The edge is one of the basic features of the image, and it is the most concentrated part of the image information. [0003] Nowadays, image edge detection technology is widely used in image cutting, image recognition and other fields, which effectively improves the accuracy and robustness of related image processing algorithms, and has great application value. In real life, image edge detect...

AI Technical Summary

Problems solved by technology

[0004] In current image edge extraction algorithms, common differential edge detection operators include Roberts operator, Prewitt operator, Sobel operator, Canny operator, etc., but these edge extraction algorithms are based on the pixel level and have high time complexity.

Images