Colony image encoding structure generation method

Abstract

The invention discloses a colony image encoding structure generation method, and mainly solves the problem that encoding efficiency is not high enough in the prior art. The scheme includes the following steps: 1) abstracting colony images into a weighted directed graph; 2) a root node passing an intermediate node, two intermediate nodes and a plurality of intermediate nodes in sequence to reach another node, continuously updating the shortest distance between the root node and each node under a depth limit, thereby obtaining a minimum spanning tree having repetitive nodes; and 3) removing the repetitive nodes through backtracking routes, obtaining a minimum spanning tree with layer restrictions, and then restoring the nodes into images, thereby obtaining the encoding structure of the colony images. Compared with the prior art, for a situation that colony images differ greatly in correlation, the encoding efficiency is higher, and the colony image encoding structure generation method can be used in scenes of image management and personal photo album management in the cloud.

Description

technical field [0001] The invention belongs to the field of image coding, and further relates to a method for generating a group image coding structure, which can be used for group image compression in the cloud, and similar scenes with image compression in professional fields, image compression in personal albums, and image database compression. Background technique [0002] Images are usually stored in JPEG format. Although it can achieve convenient access, it also has the disadvantages of large storage space and low encoding efficiency. The existing group image coding mainly utilizes the correlation among group images to generate a pseudo-video structure, and then uses video compression technology for compression. The pseudo-video structure is the group image coding structure. There are two main considerations in generating a group image coding structure: one is to improve coding efficiency, and the other is to ensure that users can read images at any time. Improve cod...

AI Technical Summary

Problems solved by technology

[0005] In his master's thesis "Research on Constrained Minimum Spanning Tree Algorithm" (Xi'an Dianzi University, 2009), he proposed a genetic algorithm to solve the diameter-constrained minimum spanning tree and a one-time construction tree method OTTC. The genetic algorithm selects The specific fitness function of , crossover, and mutation screens the directed edges, and finally obtains the smallest tree graph; the one-time tree construction method OTTC is based on the greedy idea, and selects the nearest vertex for each iteration without violating the diameter limit. Next, add this vertex, continuously update the distance between the image in the tree and the image not in the tree, and select the shortest distance accordingly, until a tree is generated, but these two methods are mainly used for undirected graphs, not suitable for Group image structure generation

[0006] In summary, the existing methods are basically not suitable for generating group image coding structures, and the HAP method, for the case where the correlation between images is greatly different, the cost of the obtained spanning tree is relatively high, and the coding efficiency is not high enough.

Images