Video Encoding Reference Frame Sampling for Cross-Resolution Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding methods encode all video frames indiscriminately, leading to poor video quality due to inefficient use of reference frames, especially under limited bandwidth conditions where frames of different resolutions are not effectively utilized.
Innovation Solution
A method and apparatus that allow for flexible selection of a sampling manner for reference frames based on the resolution of current frames, enabling sampling and reconstruction to adjust the reference frame resolution, thereby allowing frames of different resolutions to be used as references, improving inter-frame prediction accuracy and image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If all video frames are encoded indiscriminately, then the encoding process is simple, but the use efficiency of reference frames is low
Solution Approach 1:
The patent applies local quality by differentiating the treatment of reference frames based on their resolution characteristics. Instead of uniformly encoding all frames, the system identifies reference frames with resolutions different from the current frame and applies selective sampling only to those, thereby improving reference frame utilization efficiency without unnecessarily complicating the overall encoding process.
Solution Approach 2:
The patent changes the sampling parameter dynamically based on resolution matching. When a reference frame's resolution differs from the current frame, sampling is applied; when resolutions match, sampling is skipped. This parameter-based conditional approach resolves the contradiction by adapting the encoding complexity to the specific reference frame characteristics.
2Manufacturing precision
If reference frames of different resolutions are not utilized, then the encoding process is straightforward, but video quality deteriorates under limited bandwidth
Solution Approach 1:
The patent applies local quality by treating reference frames differently based on their resolution compatibility with the current frame. Reference frames with matching resolutions are used directly without sampling, while those with different resolutions undergo selective sampling. This localized differentiation improves video quality by effectively utilizing more reference frames without uniformly increasing processing complexity.
Solution Approach 2:
The patent applies partial action by performing sampling only on the subset of reference frames that require it (those with different resolutions), rather than applying sampling to all reference frames. This selective approach improves video quality by utilizing more reference frames while avoiding unnecessary processing complexity on frames that don't require sampling.
3Adaptability or versatility
If reference frames are sampled and reconstructed to adjust resolution, then reference frame usability improves, but data processing complexity increases
Solution Approach 1:
The patent applies local quality by performing sampling and reconstruction only on reference frames where resolution adjustment is actually needed. This localized application of the complex sampling operation improves reference frame usability for cross-resolution scenarios while minimizing the overall processing complexity by avoiding unnecessary sampling on resolution-matched frames.
Solution Approach 2:
The patent applies partial action by performing the complex sampling and reconstruction operation only on the partial set of reference frames that require resolution adjustment, rather than applying it to all reference frames. This selective approach enhances reference frame usability while keeping the overall system complexity manageable.
Data Source
AI summary
A video encoding method, performed by a computer device, includes: obtaining a reference frame corresponding to a current frame from a video input to be encoded; determining a sampling manner corresponding to the current frame; sampling the reference frame based on the sampling manner according to resolution information of the current frame, to obtain a target reference frame corresponding to the reference frame; and encoding the current frame according to the target reference frame.