Composite degraded image decoupling analysis and restoration method based on cross-branch connection network

Abstract

The invention discloses a composite degraded image decoupling analysis and restoration method based on a cross-branch connection network, and belongs to the field of digital image / video signal processing. A multi-branch parallel network structure is designed, a coding and decoding network structure is arranged in a branch network, short links and long links are added in the coding and decoding network structure so that the branch network can fully fuse semantic features of a low layer and a high layer, meanwhile, an attention mechanism is added in the branch network, and the attention of the low layer and the high layer can be fully fused. Therefore, the network can adaptively and dynamically adjust the network parameters according to the degradation factors of different degrees, and a certain dynamic adjustment capability is achieved. In addition, cross-branch connection is added between the branches, so that only one degradation factor is extracted from different branches, and the generation sequence of different degradation factors is also considered. And inputting the weighted fused degraded features into a reconstruction module to obtain a clear restored image. The technology has wide application prospects in the fields of criminal investigation, target tracking, military reconnaissance and the like.

Description

technical field [0001] The invention belongs to the field of digital image / video signal processing, in particular to a method for decoupling analysis and restoration of composite degraded images based on a cross-branch connection network. Background technique [0002] With the continuous development of information technology, images have increasingly become an important way for us to obtain information. However, due to the influence of shooting equipment, shooting environment, and camera shake, the images obtained in real scenes will be mixed with different types of degradation factors, such as blur, noise, compression, etc., which bring extremely bad subjective effects to the human eye. feel. Therefore, related research on complex degraded image restoration has important academic and application value. [0003] The composite degraded image collected in the real natural environment is different from the single degraded image in the laboratory environment. The degradation t...

AI Technical Summary

Problems solved by technology

For the first problem, the technology of the present invention transforms the very complex and difficult problem of finding a balance point between different degradations in most of the original methods into a relatively simple problem of finding a balance point within the same degradation, which is the volume It is difficult for the product neural network to find a balance point between different degradations, which provides a new solution

Images