Key-Frame Networks for Ultra-HD Video Quality and Bitrate
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding models suffer from video quality degradation and increased bit rates when processing high-definition and ultra-high-definition videos, leading to poor encoding and decoding effects.
Innovation Solution
A method involving extracting key and estimated frames from a sample video, performing encoding and decoding using pre-trained networks, and optimizing the model based on the resulting frames to obtain a target video encoding and decoding model, which is then used for processing target videos.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If existing video encoding and decoding models are used on high-definition and ultra high-definition videos, then encoding and decoding processing can be performed, but video quality degrades and bit rate increases
Solution Approach 1:
The patent segments the video encoding and decoding model into two distinct networks: a key frame network for processing key frames and an estimated frame network for processing estimated frames. This segmentation allows each network to be optimized for its specific frame type, improving overall video quality while managing bit rate more effectively. The key frame network focuses on high-quality reconstruction of important frames, while the estimated frame network handles intermediate frames with reduced complexity.
Solution Approach 2:
The patent applies local quality by using different encoding and decoding strategies for different frame types. Key frames, which are critical for video quality, are processed with higher precision through the key frame network. Estimated frames are processed with optimized complexity through the estimated frame network. This localized approach ensures high video quality where needed while controlling overall bit rate.
2Reliability
If existing video encoding and decoding models are used on high-definition and ultra high-definition videos, then encoding and decoding processing can be performed, but encoding and decoding effect deteriorates
Solution Approach 1:
The patent divides the encoding and decoding process into two separate networks operating in parallel. The key frame network handles key frames with optimized reliability, while the estimated frame network handles other frames with optimized efficiency. This segmentation resolves the contradiction by allowing each network to be tuned for its primary objective without compromising the other.
Solution Approach 2:
The patent introduces dynamic frame classification where frames are dynamically identified as either key frames or estimated frames based on video content characteristics. This dynamic approach allows the system to adaptively allocate processing resources, ensuring high encoding and decoding effect for important frames while maintaining efficiency for less critical frames.
3Manufacturing precision
If video data is directly transmitted without encoding and decoding, then original video quality is preserved, but large amount of network bandwidth and storage space are occupied
Solution Approach 1:
The patent extracts and processes only the essential visual information from video frames through selective encoding. By identifying key frames and estimated frames separately, the system extracts only the necessary data for high-quality reconstruction, discarding redundant information. This extraction approach maintains video quality while significantly reducing the amount of data that needs to be transmitted and stored.
Solution Approach 2:
The patent changes the parameter representation of video data by transforming visual information into compressed encoded form. The key frame network and estimated frame network apply different compression parameters and transformation techniques appropriate for each frame type, achieving efficient data reduction while preserving essential visual quality characteristics.
Data Source
AI summary
A method includes: extracting a video frame sequence from a sample video, the video frame sequence including a key frame and an estimated frame; performing encoding and decoding processing on the key frame via a pre-trained key frame network of a video encoding and decoding model, to obtain a first encoded frame and a corresponding first reconstructed frame; performing encoding and decoding processing on the estimated frame via a pre-trained estimated frame network of the video encoding and decoding model, to obtain a second encoded frame and a corresponding second reconstructed frame; performing model optimization on the video encoding and decoding model based on the first encoded frame, the first reconstructed frame, the second encoded frame, and the second reconstructed frame, to obtain a target video encoding and decoding model; and performing encoding and decoding processing on a target video by using the target video encoding and decoding model.