Image super-resolution method based on multi-scale detail feature fusion neural network

Abstract

The invention belongs to the field of image super-resolution reconstruction, and particularly relates to an image super-resolution method based on a multi-scale detail feature fusion neural network, and the method comprises the steps: obtaining a to-be-processed image, inputting the to-be-processed image into a trained improved deep neural network model, and obtaining a high-quality image; in the feature extraction process, high-frequency information in the original picture is increased, and the stability of the network is improved through the integration mode; in addition, in the aspect of a residual dense block, a multi-layer feature fusion mechanism is utilized, semantic information is added, and channel feature screening is also added, so that the whole network can be better expressed; moreover, in each stage of the feature extraction process, namely after each residual dense block, loss calculation is added for the back propagation adjustment feature extraction process, so that the expression ability of the network is improved, and high-resolution pictures can be better learned and reconstructed.

Description

technical field [0001] The invention belongs to the field of image super-resolution reconstruction, and in particular relates to an image super-resolution method based on multi-scale detail feature fusion neural network. Background technique [0002] With the rapid development of computer technology, information processing technology and visual communication technology, human beings have entered a new information age. The amount of knowledge that people can acquire is growing explosively, so it is urgent to continuously improve and develop information processing technology in order to provide people with more convenient, fast and diversified services. Digital image and its related processing technology is one of the important contents of information processing technology, and it has been more and more widely used in many fields. For digital images, high-resolution images are generally required in some cases, such as: medical images require the ability to display subtle lesi...

AI Technical Summary

Problems solved by technology

The image super-resolution task aims to reconstruct the local detail features of the image. However, the deep neural network algorithm will have different degrees of loss of high-frequency information during the feature extraction process (most of the local detail features belong to high-frequency information), making the final high-frequency information Resolution The resolution of the image does not meet the requirements

Images