Image blind super-resolution method and system

Abstract

The invention discloses an image blind super-resolution method and system. The method comprises the following steps: acquiring a trained blurring kernel generation network and a spectrogram of a low-resolution image; inputting the spectrogram into the trained blurring kernel generation network to obtain a blurring kernel corresponding to the low-resolution image; determining a degradation feature map corresponding to the low-resolution image according to the fuzzy kernel corresponding to the low-resolution image; splicing the low-resolution image and the corresponding degradation feature graph to obtain a spliced graph; and inputting the spliced image into a trained convolutional neural network to obtain a high-resolution image. The method has the advantages of simple and rapid operation and high accuracy.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image blind super-resolution method and system. Background technique [0002] Image super-resolution (Super-Resolution, SR) technology refers to given one or more low-resolution (Low-Resolution, LR) images, using a specific algorithm to reconstruct them into an accurate high-resolution (High-resolution) image. -Resolution, HR) image technology. Image super-resolution technology is widely used in many fields such as surveillance imaging, remote sensing imaging, and medical imaging, and is also used as a preprocessing algorithm for various computer vision tasks. Most image super-resolution methods assume that the HR image is obtained from the LR image through a fixed degradation method (such as bicubic interpolation and downsampling). These methods are essentially the inverse process of learning a predefined single degradation method, which cannot be flexibly To deal with multi...

AI Technical Summary

Problems solved by technology

Traditional machine learning methods usually require a large number of optimization iterations, complex calculations, long running time, and poor estimation accuracy

In the method based on deep learning, the classic fuzzy kernel estimation method KernelGAN also has the problem of complex operation and long running time.

Images