Mobile equipment panoramic video playing system based on deep reinforcement learning
A panoramic video and reinforcement learning technology, applied in image communication, selective content distribution, electrical components, etc., can solve the problems of inconvenient use, inconvenient carrying, heavy weight, etc., to reduce bandwidth consumption, avoid video stagnation, and smooth The effect of playing
Inactive Publication Date: 2019-03-01
SUN YAT SEN UNIV
View PDF7 Cites 13 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Using wired head-mounted devices for panoramic video playback, because head-mounted devices are relatively heavier, inconvenient to carry, and more expensive, they can only be arranged in corresponding places, which is not convenient for daily use. The solution for wireless mobile device transmission is generally to use LTE or WIFI to transmit panoramic video
[0003] However, due to the high requirements for panoramic video transmission on the network, the current LTE or WIFI cannot meet the needs of high-quality panoramic video transmission, so relatively speaking, there are almost no panoramic video transmission solutions for wireless mobile devices on the market.
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0042] The code rate selection process of the code rate selection module is similar to the Markov decision process, that is, the next action and state are only determined by the current state. The deep reinforcement learning method is used to solve this problem. The network conditions, the file size to be downloaded for each version definition of the next time slice, the video buffer conditions, and the motion speed are defined as the state, and the bit rate selection of each panoramic video block is defined as the action, the user experience obtained, and the buffer The occupancy situation is defined as reward, build a reinforcement learning model, and then put it into training, and finally get a bit rate selection module, which can dynamically select the next moment according to the current state, and present the video block that needs to be obtained for each block of the video clarity version.
[0043] The specific playback process of the panoramic video is as follows. Firs...
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
Login to View More Abstract
The invention discloses a mobile equipment panoramic video playing system based on deep reinforcement learning, which utilizes three technologies consisting of panoramic video block transmission, dynamic rate selection and motion prediction to provide experience of playing a panoramic video on wireless mobile equipment for a user while ensuring video quality and saving network bandwidth. The mobile equipment panoramic video playing system of the invention has the beneficial effects that: panoramic video information is divided into a plurality of panoramic video blocks with a plurality of definitions, a deep reinforcement learning method is adopted in the dynamic rate selection to automatically select a rate and automatically predict an effect of head movement according to the motion prediction, panoramic video definitions can be selected according to viewing angle changes of the users, and finally a smooth panoramic video with different definitions can be displayed in a video playing module, so as to bring good experience to the users.
Description
technical field [0001] The present invention relates to the technical field of computer software, and more specifically relates to a mobile device panoramic video playback system based on deep reinforcement learning. Background technique [0002] Panoramic video is a form of virtual reality, that is, VR. At present, panoramic video is more and more used in daily life. The means of panoramic video transmission are mostly presented through wired head-mounted devices. This is because the current Technology, only through wired equipment can the bandwidth requirements for video image transmission be met. Using wired head-mounted devices for panoramic video playback, because head-mounted devices are relatively heavier, inconvenient to carry, and more expensive, so they can only be arranged in corresponding places, which is not convenient for daily use. The solution for wireless mobile device transmission is generally to use LTE or WIFI to transmit panoramic video. [0003] Howev...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): H04N21/2343H04N21/262H04N21/4363H04N21/6373H04N21/81
CPCH04N21/2343H04N21/234309H04N21/26208H04N21/43637H04N21/6373H04N21/8146
Inventor 肖功伟陈旭
Owner SUN YAT SEN UNIV