Recursive residual network-based super-resolution image reconstruction method

Abstract

Embodiments of the invention provide a recursive residual network-based super-resolution image reconstruction method. The method comprises the following step of: inputting a low-resolution image intoa trained recursive residual neural network so as to obtain a super-resolution reconstructed image, wherein the recursive residual neural network comprises a plurality of residual units, and for any residual unit, input information of the residual unit comprises output information of the last residual unit and a high-frequency feature image of a low-resolution input image. According to the method,the neural network is trained through local residual learning but not global residual learning of VDSR, so that benefit is brought to information transmission and gradient flow; non-interpolated low-resolution images are taken as inputs; and finally, super-resolution output images are directly up-sampled by using a convolution layer at the tail end of the network, and recursive structures are imported in the residual units, so that parameters are greatly decreased and the calculation complexity of the recursive residual neural network is reduced.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of image reconstruction, and in particular to a super-resolution image reconstruction method based on a recursive residual network. Background technique [0002] Single image super-resolution (Single image super-resolution, referred to as SISR) is a classic computer vision problem, aiming at recovering high-resolution (High-resolution) from a given low-resolution (Low-resolution, referred to as LR) image. -resolution, referred to as HR) image. Because SISR recovers high-frequency information, it is widely used in fields that require more image details, such as medical imaging, satellite imaging, security monitoring, etc. [0003] Existing super-resolution (SR) image reconstruction methods are mainly divided into three categories: interpolation-based SR technology, reconstruction-based SR technology and learning-based SR technology. Most of the current SR algorithms are learning-based ...

AI Technical Summary

Problems solved by technology

In practical applications, many such as mobile devices and embedded devices are limited in terms of calculation, volume, power consumption, etc., resulting in the inability of existing high-performance deep neural networks to perform effective calculations and applications on them.

Images