Image restoration method and device

Abstract

The invention provides an image restoration method and device, and the method comprises the steps: obtaining a to-be-restored image, and carrying out the distinguishing of the to-be-restored image, soas to obtain a to-be-restored region and a remaining region; calculating the autocorrelation degree of the to-be-repaired area and the remaining area, and taking the area with the highest autocorrelation degree as a reference area; performing high-frequency filtering on the image in the reference area by adopting a high-frequency filtering algorithm to obtain high-frequency information; inputtingthe images corresponding to the remaining areas into a pre-trained neural network model to output low-frequency information; mixing the low-frequency information with the high-frequency information to obtain to-be-filled information; filling the to-be-repaired area with the to-be-filled information so as to repair the to-be-repaired image. Therefore, the low-frequency information is acquired through the deep learning network, and the high-frequency information is acquired through the intelligent filling method, so that the filling result has both the continuity of the peripheral information and the authenticity of the information, and the image restoration effect is greatly improved.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image restoration method, an image restoration device, a computer-readable storage medium and computer equipment. Background technique [0002] Pictures can record the details of life, but due to the influence of shooting conditions and techniques, the pictures often need to be processed, such as erasing redundant ornaments or decorations on the background, and after erasing unwanted content , it is also necessary to restore the image, that is, to fill the erased area with new content information, so that the filled new information can have both the continuity of the surrounding information and the authenticity of the information itself. [0003] There are usually two commonly used image repair methods at present. One is to use the intelligent filling function of WebInpaint or PhotoShop to directly cover the content information around the area to be filled on ...

AI Technical Summary

Problems solved by technology

[0003] There are usually two commonly used image repair methods at present. One is to use the intelligent filling function of WebInpaint or PhotoShop to directly cover the content information around the area to be filled on the area to be filled. Although this method can be better Realize the authenticity of the filling effect, but the continuity between the filling information and the original surrounding information is very poor; the other is to use deep learning technology to train large data samples, so that the deep neural network can automatically generate data according to the surrounding information form. The ability of information, this method can realize the continuity of filling information and the original surrounding information, but lacks authenticity

That is to say, the existing image restoration methods can not take into account the continuity of surrounding information and the authenticity of the information itself, resulting in poor restoration effect.

Images