Video restoration method, device and equipment and storage medium

A restoration method and video technology, applied in image data processing, instrumentation, computing, etc., can solve problems such as reverse effects of video restoration, affecting the quality of video restoration, etc., to achieve the effect of improving the effect, improving versatility, and improving video quality

Active Publication Date: 2020-12-18
TENCENT TECH (SHENZHEN) CO LTD
View PDF7 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method relies on the prediction of degradation intensity. If the prediction is inaccurate, it will lead to the opposite effect of subsequent video restoration, thus affecting the quality of video restoration.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Video restoration method, device and equipment and storage medium
  • Video restoration method, device and equipment and storage medium
  • Video restoration method, device and equipment and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach 1

[0107] Embodiment 1. The video encoding parameters include a macroblock reference (Macroblock Tree, MBtree), and the types of areas to be compensated include static areas and dynamic areas. An area in the video frame to be repaired whose macroblock reference value is greater than the first threshold is determined as a static area, and an area in the video frame to be repaired whose macroblock reference value is not greater than the first threshold is determined as a dynamic area.

[0108] Specifically, MBtree is a quantization parameter control method based on macroblocks. When the value of MBtree is larger, it means that the corresponding area is more likely to be a slowly changing scene; when the value of MBtree is smaller, it is more likely that the corresponding area is A scene of drastic change. When the video frame to be repaired is divided into multiple regions to be compensated, the distribution image of the macro reference value corresponding to the video frame to be ...

Embodiment approach 2

[0109] Embodiment 2. The video coding parameters include adaptive quantization (Adaptive Quantization, AQ), and the types of regions to be compensated include texture regions and flat regions. Determining an area in the video frame to be repaired whose adaptive quantization value is greater than the second threshold as a texture area, and determining an area in the video frame to be repaired whose adaptive quantization value is not greater than the second threshold as a flat area.

[0110] Specifically, when the value of AQ is larger, it indicates that the degree of detail of the corresponding region is high, and when the value of AQ is smaller, it indicates that the degree of detail of the corresponding region is low. When the video frame to be repaired is divided into multiple regions to be compensated, the distribution image of the adaptive quantization value corresponding to the video frame to be repaired can be obtained first. The image content of different regions in the ...

Embodiment approach 3

[0111] Embodiment 3: The video encoding parameters include macroblock reference and adaptive quantization, and the types of regions to be compensated include static regions, dynamic regions, texture regions and flat regions. Determining an area in the video frame to be repaired whose macroblock reference value is greater than a first threshold as a static area; determining an area in the video frame to be repaired whose macroblock reference value is not greater than the first threshold as a dynamic area. Determining an area in the video frame to be repaired whose adaptive quantization value is greater than a second threshold as a texture area; determining an area in the video frame to be repaired whose adaptive quantization value is not greater than the second threshold as a flat area.

[0112] In the third embodiment, there may be a case where the area to be compensated overlaps, for example, there is a partial overlapping area between the static area and the flat area. It sh...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The embodiment of the invention provides a video restoration method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method specifically comprises the steps: employing a neural network model to extract the image features of a to-be-restored video frame, and predicting the image compensation parameter value of the to-be-restored video frame based on the image features. And dividing the to-be-restored video frame into a plurality of to-be-compensated regions according to the distribution characteristics of the video coding parameters corresponding to the to-be-restored video frame. And determining a region compensation parameter value of each region to be compensated according to the type of each region to be compensated, a preset image compensation strategy and the image compensation parameter value. And performing image compensation on each to-be-compensated region based on the region compensation parameter value of eachto-be-compensated region to obtain a restored video frame. Dividing the to-be-restored video frame into the plurality of to-be-compensated regions, determining the region compensation parameter valueof each to-be-compensated region based on the predicted image compensation parameter value, and carrying out targeted image compensation, so that the video restoration effect is improved.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of artificial intelligence, and in particular to a video restoration method, device, equipment and storage medium. Background technique [0002] At present, video encoding will cause compression distortion. Common types of distortion include blocking effects, mosaics, edge glitches, and ringing effects. These compression noises have a great impact on the viewer's visual experience. In the related art, different decompression and distortion processing methods are used to repair videos with different degradation degrees. For example, the encoding parameters of QuantizationParameter (QP) 20, QP30, and QP40 are compressed to generate degraded images with different degraded intensities, and then the degraded images with different degraded intensities are used to train degradation models of different levels. When decompressing and distorting a video, the degradation intensity of the video i...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T5/00
CPCG06T5/005G06T2207/10016G06T2207/20081G06T2207/20084
Inventor 江林燕李雅卿涂承杰彭龙涛熊诗尧王茹
Owner TENCENT TECH (SHENZHEN) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap