20230105436. GENERATIVE ADVERSARIAL NETWORK FOR VIDEO COMPRESSION simplified abstract (KWAI INC.)
Contents
GENERATIVE ADVERSARIAL NETWORK FOR VIDEO COMPRESSION
Organization Name
Inventor(s)
Ying Liu of Santa Clara CA (US)
Lingzhi Liu of San Jose CA (US)
Yongxiong Ren of San Jose CA (US)
Ming Kai Hsu of Fremont CA (US)
GENERATIVE ADVERSARIAL NETWORK FOR VIDEO COMPRESSION - A simplified explanation of the abstract
This abstract first appeared for US patent application 20230105436 titled 'GENERATIVE ADVERSARIAL NETWORK FOR VIDEO COMPRESSION
Simplified Explanation
The abstract describes a method and apparatus for video processing using generative adversarial networks (GANs). The method involves a decoding terminal receiving coded video frames and network parameters related to the GANs. The terminal then decodes the coded video frames using GANs based on the network parameters. The GANs implement various video coding functions such as reference-frame coding, motion-compensated frame prediction, and residue-frame coding.
- The method involves using generative adversarial networks (GANs) for video processing.
- A decoding terminal receives coded video frames and network parameters related to the GANs.
- The coded video frames are decoded using GANs based on the network parameters.
- The GANs implement video coding functions including reference-frame coding, motion-compensated frame prediction, and residue-frame coding.
Potential applications of this technology:
- Video compression and encoding: The use of GANs for video processing can improve video compression and encoding techniques, leading to more efficient storage and transmission of video data.
- Video streaming: By utilizing GANs for video processing, video streaming services can enhance the quality and reduce the bandwidth requirements for streaming video content.
- Video editing and post-production: GANs can be used to enhance video editing and post-production processes by improving the quality and visual effects of the edited video.
Problems solved by this technology:
- Inefficient video coding: Traditional video coding techniques may not fully utilize the potential of GANs for improved video compression and encoding.
- Bandwidth limitations: GAN-based video processing can help overcome bandwidth limitations by reducing the amount of data required for video transmission.
- Quality degradation: By implementing GANs for video coding functions, the quality of decoded video frames can be improved, reducing artifacts and enhancing visual fidelity.
Benefits of this technology:
- Improved video compression: GAN-based video processing can lead to more efficient video compression techniques, resulting in reduced storage requirements and faster transmission.
- Enhanced video quality: By utilizing GANs for video coding functions, the quality of decoded video frames can be improved, providing a better viewing experience for users.
- Bandwidth optimization: GAN-based video processing can help optimize bandwidth usage, allowing for smoother video streaming and reduced buffering times.
Original Abstract Submitted
a method and an apparatus for video processing are provided. the method includes that a decoding terminal receives a plurality of coded video frames coded using one or more generative adversarial networks (gans), receives network parameters related to the one or more gans, and decodes the plurality of coded video frames using gans based on the network parameters. further, the one or more gans respectively implement one or more video coding functions including reference-frame coding, motion-compensated frame prediction, and residue-frame coding.
