Multi-traversal video coding rate allocation and control optimization method based on reinforcement learning

A technology of reinforcement learning and video coding, applied in the fields of video coding and deep learning, can solve the problems such as the inability to obtain the global optimal solution and the difficulty of generating labels with machine learning methods, achieve practical significance and industrial value, improve coding and compression efficiency, The effect of improving compression efficiency

Active Publication Date: 2021-04-23
杭州微帧信息科技有限公司
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The invention solves the problem that the traditional method cannot obtain the global optimal solution, and the machine learning method is difficult to generate labels for training

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
  • Multi-traversal video coding rate allocation and control optimization method based on reinforcement learning
  • Multi-traversal video coding rate allocation and control optimization method based on reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] The following examples will be used in conjunction with the accompanying drawings. The following examples are only used to illustrate the technical solution of the present invention more clearly, but not to limit the protection scope of the present invention.

[0023] The optimization method of code rate allocation and control in video multiple encoding of the present invention specifically optimizes the code control allocation and control strategy parameters of the second traversal of video encoding, including the following steps,

[0024] Step (1), create prediction network and discriminant network, and complete network parameter initialization.

[0025] The prediction network is a fully connected neural network, which is responsible for determining the code rate allocation in the picture group and controlling the corresponding code rate allocation and control strategy parameter optimal value based on the comprehensive statistical information generated by the first tra...

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 an optimization solution for code rate allocation and control in multipass video encoding based on reinforcement learning. After the first encoding, the video is divided into multiple non-overlapping continuous picture groups. According to the comprehensive statistical information of each frame image in the group collected after the first traversal (pass) of different picture groups, the code rate allocation and control decision parameters are obtained through the prediction network, and the second traversal encoding is performed and the encoding result is obtained score. Input the statistical information and code control decision parameters into the discriminant network to obtain the estimated score, and perform the iterative training process of reinforcement learning. Using the prediction network obtained by reinforcement learning training, before the second traversal of the encoded video image, the optimal strategy parameters for the bit rate allocation and control of each picture group are obtained, so as to improve the encoding and compression efficiency as much as possible.

Description

technical field [0001] The invention relates to video coding and deep learning, in particular to an optimization method for multi-traversal video coding code rate allocation and control based on reinforcement learning. Background technique [0002] With the continuous development of multimedia digital video applications and the continuous improvement of people's demand for video cloud computing, the data volume of the original video source makes the existing transmission network bandwidth and storage resources unbearable. Therefore, the compression of video signal has become one of the hot spots of academic research and industrial application at home and abroad. Video compression, also known as video coding, aims to eliminate redundant information between video signals. So far, domestic and foreign standardization organizations have successively formulated a variety of different video coding standards. Since the H.261 video coding standard, the mainstream video coding stan...

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 Patents(China)
IPC IPC(8): H04N19/115H04N19/146
CPCH04N19/115H04N19/146
Inventor 朱政陈宇梅元刚丁丹丹
Owner 杭州微帧信息科技有限公司
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