Picture high-definition reconstruction method, device, mobile terminal and storage medium

Abstract

The embodiment of the invention discloses a picture high-definition reconstruction method, device, mobile terminal and storage medium. The method includes performing combined dictionary training on low-definition sample pictures and high-definition sample pictures by adopting sparse coding and obtaining an overcomplete high-low dictionary pair, wherein the overcomplete high-low dictionary pair includes a low-definition dictionary and high-definition dictionary; obtaining a sparse representation coefficient of a low-resolution image to be reconstructed by utilizing the low-definition dictionary; and obtaining a reconstructed high-resolution image of the low-resolution image to be reconstructed by utilizing the sparse representation coefficient and the high-definition dictionary. Dictionarytraining time can be shortened and the overcomplete high-low dictionary pair for universal application and having clear structure can be acquired without need of an excessive large sample set. Besides, the resolution rate of the reconstructed high-resolution image can be increased effectively.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image super-resolution reconstruction method, device, mobile terminal and storage medium. Background technique [0002] In the prior art, the image super-resolution algorithm is mainly divided into two categories, one is the image super-resolution algorithm based on the reconstruction algorithm, and the other is the image super-resolution algorithm based on the learning algorithm. Among them, the existing reconstruction-based The image super-resolution algorithm of the algorithm does not improve the resolution of the image significantly, especially for images with complex structures, the restored image is still relatively blurred; the existing image super-resolution algorithm based on learning algorithms requires The sample set is too large, the training time is too long, and the trained dictionary cannot be universally applicable. [0003] Therefore, the exi...

AI Technical Summary

Problems solved by technology

[0004] The main purpose of the embodiments of the present invention is to provide an image super-resolution reconstruction method, device, mobile terminal and storage medium, which solves the problem that the image resolution improvement in the existing image super-resolution algorithm is not obvious and needs The technical problems that the sample set is too large, the training time is too long, and the trained dictionary cannot be universally applicable

Images