Super-resolution reconstruction method based on dense neural network and two-parameter loss function

Abstract

The invention discloses a super-resolution reconstruction method based on a dense neural network and a two-parameter loss function. The method comprises a data preparation stage, a network structure construction stage, a model training stage, and an image reconstruction stage. The idea of a dense convolutional neural network structure is applied to the super-resolution reconstruction of a single-frame image, shallow features such as texture and contour in the network structure are fully utilized, and local features and global features of the image are combined to perform super-resolution reconstruction. On the above basis, the two-parameter loss function is adopted to perform more accurate optimization training on the network structure, and the algorithm effect is improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and more specifically relates to a super-resolution reconstruction method based on a dense neural network and a double-parameter loss function. Background technique [0002] The super-resolution reconstruction of the image is to use various reconstruction techniques to map the low-resolution image into a high-resolution image, recover more high-frequency information, and make the texture and texture of the image clearer. Since multiple different high-resolution images may produce the same low-resolution image through the same downsampling, which is a typical one-to-many problem, the image super-resolution reconstruction problem is essentially an ill-posed problem. )question. According to the different principles of the algorithm, image super-resolution reconstruction methods can be roughly divided into three categories: interpolation-based methods, reconstruction-based methods, and lear...

AI Technical Summary

Problems solved by technology

However, the L2 norm makes the reconstructed image too smooth due to the square term

Compared with the L2 norm, although the L1 norm can make the values ​​tend to be sparse during the training process, it still cannot reconstruct images with sufficient high-frequency information.

Images