Content-Aware DNN for Adaptive Video Streaming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video streaming services face quality degradation due to dependence on transmission bandwidth, and existing solutions like distributed computing and adaptive bitrate streaming are not completely independent of bandwidth variations.
Innovation Solution
The use of content-aware deep neural networks (DNNs) by both server and client to encode, train, and transmit adaptive video, allowing for real-time quality enhancement and optimization based on client hardware resources, reducing bandwidth requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If adaptive bitrate streaming and distributed computing are used to improve video quality, then video quality is improved, but the system remains dependent on transmission bandwidth variations
Solution Approach 1:
The patent replaces traditional mechanical bandwidth-dependent video processing with a neural network-based system. The content-aware DNN performs video enhancement tasks (super-resolution, de-blocking, de-interlacing) that traditionally required high-bandwidth transmission of high-quality video, thereby reducing bandwidth dependency while maintaining video quality
Solution Approach 2:
The system dynamically changes video processing parameters by selecting different pre-trained DNN models based on network conditions and client capabilities. The server can switch between different enhancement models (e.g., super-resolution vs. de-blocking) to optimize video quality while adapting to varying bandwidth conditions
2Manufacturing precision
If content-aware DNNs are trained and transmitted to enhance video quality, then video quality is improved, but transmission bandwidth is reduced
Solution Approach 1:
The server pre-trains content-aware DNN models for various video enhancement tasks before transmission. These pre-trained models are packaged in configuration files and transmitted to clients, where they perform real-time video enhancement without requiring additional bandwidth for processing high-quality video streams
Solution Approach 2:
Instead of transmitting multiple versions of video at different qualities, the system transmits a single video stream along with copied DNN models that enable the client to enhance the video locally. The DNN models act as virtual copies of server-side processing capabilities
3Manufacturing precision
If video is encoded at multiple resolutions and transmitted, then video quality is improved, but the complexity of the system increases
Solution Approach 1:
The patent extracts the video enhancement functionality from the encoding process. Instead of encoding videos at multiple resolutions during the encoding phase, the system extracts a single-resolution video and uses separate DNN models to enhance it to different quality levels, thereby simplifying the encoding complexity
Data Source
AI summary
A method and apparatus for transmitting adaptive video in real time using a content-aware neural network are disclosed. At least one embodiment provides a method performed by a server for transmitting an adaptive video in real time by using content-aware deep neural networks (DNNs), including downloading a video, encoding a downloaded video for each of at least one resolution, dividing an encoded video into video chunks of a predetermined size, training the content-aware DNNs by using encoded video, generating a configuration or manifest file containing information on trained content-aware DNNs and information on the encoded video, and transmitting the configuration file upon a request of a client.