Image restoration method based on wavelet transform attention model

Abstract

The invention relates to an image restoration method based on a wavelet transform attention model, and is used for providing a high-quality image restoration method. The method is based on an image restoration network, and the network is based on a GAN and comprises a generator network and a discriminator network. The generator network part firstly utilizes DWT to decompose known region features in a damaged image into multi-frequency sub-bands, then further extracts deep information of a known region through an attention mechanism, and finally generates a restored image through IDWT. A full convolution discriminator network structure of a PatchGAN is adopted in a discriminator network part; the restored image and the ground-true image are jointly used as input, the generator network and the discriminator network are alternately optimized, back propagation is carried out by minimizing a global loss function, generator network parameters are iteratively adjusted, and finally, the optimized generator network can generate a high-quality image restoration result.

Description

Technical field: [0001] The invention relates to the field of computer image processing, in particular to an image restoration method based on a wavelet transform attention model and a generation confrontation network. Background technique: [0002] Image inpainting refers to the technology of inferring the content of the missing area and repairing it by using a certain method to infer the content of the missing area through the known area information of the image. This is a classic and challenging image processing topic widely used in image editing, image-based rendering, and computational photography. There are two problems to be solved in the image restoration work: first, in the restoration process, the continuity of the texture structure of the image should be maintained, so that the missing area should be as close as possible to or achieve the effect of the original image; second, the restored image should be real Natural, in line with people's visual coherence, let t...

AI Technical Summary

Problems solved by technology

(H.Ishikawa, H.Ishikawa, and H.Ishikawa, Globally and locally consistent image completion, 2017.) Although the image restoration method based on convolutional neural network can deal with complex scenes or image restoration problems with many missing areas, it usually It will cause problems such as structural distortion, blurred texture and incoherent content with obvious boundaries in the repaired image

[0004] In summary, the existing image inpainting methods repair images with low quality, especially for images with complex texture structures or large missing areas, which may easily cause problems such as structural distortion, blurred texture details, and incoherent content in the repair results. certain limitations

Images