Super-resolution image reconstruction method and system based on multi-feature learning

Abstract

The invention discloses a super-resolution image reconstruction method and system based on multi-feature learning, and the method makes full use of the rich information contained in a single input image for reconstruction, and does not depend on an external database. According to the method, a mapping relation between image features is established based on cross-scale similarity of images, and a high-resolution image containing high-frequency information is reconstructed for an input image directly by using the mapping relation, so that the defect of high-frequency information loss caused by image reconstruction by using an interpolation amplification method is well overcome. According to the method, effective high-frequency information is acquired by using singular value thresholding, andthe high-frequency information is amplified by using a gradient feature mapping relation and then is overlapped on a high-resolution image in a blocking manner, so that a final image reconstruction result is obtained. According to the method for reconstructing the image by utilizing the image feature combination, noise points of the reconstructed image are effectively inhibited, image edge and texture information is well kept, and detail enhancement of the image is realized.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, and in particular, to a method and system for super-resolution image reconstruction based on multi-feature learning. Background technique [0002] The statements in this section merely enhance the background related to the present disclosure and do not necessarily constitute prior art. [0003] Image super-resolution technology is to convert a low resolution image (Low Resolution Image, LR) into a high resolution image (High Resolution Image, HR) by certain means. In this process, the effective information of the image is preserved as much as possible, including the structure, texture, details, etc. of the image. [0004] At present, there are many research methods of image super-resolution technology, which are mainly divided into interpolation-based super-resolution methods, reconstruction-based super-resolution methods and learning-based super-resolution methods. [0005] T...

AI Technical Summary

Problems solved by technology

Images reconstructed by this type of method are prone to blurred edges and poor overall visual effects.

In order to improve the defects caused by traditional interpolation methods, many new interpolation methods have emerged. Although these methods effectively overcome the defects of traditional interpolation methods, it is still difficult to achieve satisfactory visual effects. For example, the edge-based image interpolation method in the image Deformation and other phenomena appear in the details and the algorithm complexity is high

[0006] The second category: reconstruction-based methods, the reconstruction results of this type of method often lack image high-frequency information, and the results often rely on reconstruction constraints, so that the image looks smoother

However, the algorithm will produce blocks in high-frequency information areas such as image details and textures.

This is because the high-frequency subbands searched by the algorithm contain pseudo-pixels, and the pseudo-pixels are placed into the result image during the superposition process, resulting in blocky images in the reconstructed result image.

Images