Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Video coding convolution filtering method based on attention mechanism fusion unit division

A technology of video coding and convolution filtering, applied in the field of video coding, can solve problems such as differences in coding efficiency, and achieve the effect of being conducive to recovery

Active Publication Date: 2021-01-22
UNIV OF ELECTRONIC SCI & TECH OF CHINA
View PDF5 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Get the coding efficiency difference of the two algorithms

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 coding convolution filtering method based on attention mechanism fusion unit division
  • Video coding convolution filtering method based on attention mechanism fusion unit division
  • Video coding convolution filtering method based on attention mechanism fusion unit division

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0060]The present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0061]The current video coding technology has losses in the quantization stage. Therefore, after the image is reconstructed, an in-loop filter is added to process the reconstructed image to improve the image quality. Common in-loop filters include deblocking filters, sample adaptive offset filters, and adaptive loop filters. Although these filters based on traditional algorithms have certain effects, they are far from meeting people's needs for video quality. In view of the advantages of the convolutional neural network for image processing, the present invention proposes a convolutional neural network structure to replace the traditional filter, and achieve a good filtering result.

[0062]Regardless of HEVC or the latest VVC, the block-based video coding technology used and the loss of information in the quantization process will cause artifacts that seriously...

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 invention discloses a video coding convolutional filtering method based on attention mechanism fusion unit division. The method comprises the steps: carrying out the downsampling of different scales of a constructed coding unit division structure diagram, extracting the feature information of receptive fields of different scales, fusing all feature information, and carrying out the processingof a multi-layer convolutional neural network layer; performing feature extraction processing on the reconstructed image through a deep convolutional neural network; taking the feature map of the coding unit division structure as a weight, weighting the weight to a feature map generated by the reconstructed image based on pixels, and enabling the network to pay more attention to the region of interest; and finally, performing multi-convolution-layer processing to obtain a final residual image, and adding the input reconstructed image to obtain a restored reconstructed image. According to the method, the coding unit structure feature information is fully extracted by using a multi-scale network, the coding unit division structure is effectively fused with the reconstructed image features, and the image quality recovery process is guided, so that the reconstructed image quality and the coding efficiency are improved.

Description

Technical field[0001]The invention relates to the technical field of video coding, in particular to a video coding convolution filtering method divided by an attention mechanism fusion unit.Background technique[0002]With the advancement of video coding technology, people have proposed many video coding standards, such as AVC (Advanced Video Coding), HEVC (High Efficiency Video Coding) and the latest VVC (Versatile Video Coding) . Compared with the previous generation of standards, the bit rate of each new standard has been reduced by 50% under the same image quality. These standards, including the latest VVC, are based on block coding. First, a frame of image is divided into coding tree units of the same size, and then further divided into coding units of appropriate size according to rate-distortion optimization. On the basis of the coding unit, the prediction unit block is obtained through intra-frame prediction or inter-frame prediction. Then subtract the prediction block from th...

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
IPC IPC(8): H04N19/82H04N19/86
CPCH04N19/82H04N19/86
Inventor 高艳博贾梦虎李帅岳建
Owner UNIV OF ELECTRONIC SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com