Video Coding Interprediction Using Adaptive Resolution Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression technologies face challenges in managing resolution efficiently, leading to increased bitrates and quality issues due to the requirement of re-coding and re-sending whole groups of pictures, which can incur additional costs and result in visible coding artifacts.
Innovation Solution
Adaptive Resolution Management (ARM) allows for prediction using reference frames of different resolutions, enabling bitrate savings by downsampling or upsampling video resolution, and eliminating the need for intra-frame re-coding, through implicit signaling based on frame type, which disallows ARM for frames used as reference frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If adaptive resolution management is applied to all frames including reference frames, then bitrate is reduced and compression efficiency is improved, but prediction accuracy deteriorates and coding artifacts increase
Solution Approach 1:
The patent applies different resolution management strategies to different frame types. Reference frames (I-frames and P-frames) are maintained at full resolution to ensure prediction accuracy, while B-frames and non-reference frames utilize adaptive resolution management with downsampling. This local differentiation ensures that critical prediction operations use high-quality full-resolution data while non-critical frames benefit from reduced bitrate through resolution adaptation.
2Productivity
If resolution changes are applied frequently within group of pictures, then compression efficiency is improved, but coding artifacts and prediction mismatches increase
Solution Approach 1:
The patent implements resolution changes periodically at specific frame boundaries rather than continuously. Resolution adaptation is applied at B-frame boundaries and non-reference frame boundaries, while reference frames maintain full resolution. This periodic application of resolution management reduces prediction mismatches and coding artifacts by ensuring that resolution changes occur only at appropriate intervals where they will not disrupt ongoing prediction operations.
3Manufacturing precision
If intra-frame re-coding is performed to maintain quality, then video quality is maintained, but bitrate increases and encoding complexity increases
Solution Approach 1:
The patent changes the resolution parameter adaptively for different frame types and positions within the group of pictures. Instead of maintaining constant full resolution for all frames, the system dynamically adjusts resolution based on frame type (I, P, B frames) and position within the GOP structure. This parameter change allows significant bitrate reduction while maintaining perceived video quality by preserving full resolution only where necessary for prediction accuracy.
4Quantity of substance
If adaptive resolution management is allowed for reference frames, then bitrate is reduced, but decoding complexity and synchronization issues increase
Solution Approach 1:
The patent segments the video stream into different frame type categories (reference frames vs. non-reference frames) and applies different resolution management rules to each segment. Reference frames are consistently maintained at full resolution, while non-reference frames utilize adaptive resolution management. This segmentation simplifies decoding by providing clear, consistent rules for different frame types, avoiding the complexity of dynamic resolution adjustment for every frame while still achieving bitrate reduction.
Data Source
AI summary
A video encoder is provided which generates a bitstream for decoding by a compliant decoder having an interprediction mode for generating a predictor from a reference picture for a subsequent picture having a spatial resolution different from the reference picture. The encoder provides a bitstream in which the decoder can scale the spatial resolution of a block of pixels of the reference picture using a scale factor signaled in the encoded bitstream and using a filter chosen for the particular block of the reference picture and not the reference picture as a whole to form a predictor.