Video restoration method and device, electronic device and storage medium

Abstract

The invention relates to a video restoration method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining a first forward optical flow and a first reverse optical flow between adjacent images in multiple continuous frames of images; performing optical flow optimization processing on the first forward optical flow and the first reverse optical flow respectively to obtain a second forward optical flow corresponding to each first forward optical flow and a second reverse optical flow corresponding to each first reverse optical flow; carrying out forward conduction optimization on the continuous multi-frame images by utilizing the second forward optical streams, and carrying out reverse conduction optimization on the continuous multi-frame imagesby utilizing the second reverse optical streams until all the images in the optimized continuous multi-frame images meet repair requirements; and according to the optimized images obtained through the forward conduction optimization and the reverse conduction optimization, obtaining the continuous restored image of the multi-frame image. According to the embodiment of the invention, the consistency of the continuous frame repair and the real-time performance of repair can be improved.

Description

technical field [0001] The present disclosure relates to the field of video restoration, and in particular to a video restoration method and device, electronic equipment and a storage medium. Background technique [0002] Image inpainting has always been an important issue in computer vision, and has important applications in many fields. For example, image dewatermarking and image inpainting can be achieved through image inpainting. Video inpainting, as an extension of image inpainting on video, is much more difficult to implement than pure image inpainting. However, with the development of the Internet, the application of video repair algorithms has a wider prospect, such as the need to remove video watermarks and certain parts of videos. [0003] At present, some existing methods for image restoration can be directly applied to video, but due to the lack of timing consistency, it is difficult to achieve continuous consistency between frames in video. Other traditional v...

AI Technical Summary

Problems solved by technology

[0003] At present, some existing methods for image restoration can be directly applied to video, but due to the lack of timing consistency, it is difficult to achieve continuous consistency between frames in video.

Other traditional video repair algorithms based on optimization, because they need to continuously search and match between image blocks, it is difficult to achieve real-time repair. For a 90-frame video, it even takes 3-4 hours. processing time

[0004] Therefore, there are technical problems in the prior art that video restoration cannot achieve continuous consistency between frames, and poor real-time performance

Images