Image restoration method and device and electronic equipment

Abstract

The embodiments of the invention provide an image restoration method and device and electronic equipment. The method comprises the following steps: carrying out partial convolution processing on a to-be-restored image by utilizing a convolution kernel with a preset size to obtain a feature map of the to-be-restored image; downsampling the feature map of the to-be-restored image to a first resolution through partial convolution processing: carrying out the partial convolution processing of the feature map of the to-be-restored image through employing the plurality of cavity convolution kernelswith different sizes, and obtaining feature maps corresponding to the plurality of cavity convolution kernels with different sizes; fusing the feature maps corresponding to all the cavity convolutionkernels of different sizes to obtain a fusion result, and taking the fusion result as a new feature map of the to-be-restored image; and through partial convolution processing, downsampling the feature map of the to-be-restored image to a second resolution: performing deconvolution processing on the feature map of the to-be-restored image to obtain an image with the same resolution as the to-be-restored image, and taking the image as a restored image. The restored image can be more accurate.

Description

technical field [0001] The present invention relates to the technical field of deep learning, in particular to an image restoration method, device and electronic equipment. Background technique [0002] Part of the image information may be missing in some images. For example, due to the occlusion of billboards, the images of some scenes are missing in the captured landscape photos, which affects the visual effect of the image. In related technologies, a neural network obtained by deep learning can be used to restore missing image information in an image. [0003] However, related methods of image restoration in the prior art are usually a process of adding information to an image based on existing information in the current image. Therefore, if too much information is missing in the current image, it is difficult to effectively extract the semantic information of the image by existing methods, and it may not be possible to perform image restoration well. Contents of the i...

AI Technical Summary

Problems solved by technology

Therefore, if too much information is missing in the current image, it is difficult to effectively extract the semantic information of the image with existing methods, and it may not be possible to perform image restoration well.

Images