Long-Term Reference Frame Retention for Video Compression Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression standards like H.264 and H.265 incur efficiency losses when scenes with static backgrounds change, as updating entire long-term reference frames is costly and inefficient, leading to repeated encoding of previously discarded frames.
Innovation Solution
Implementing extended long-term reference (eLTR) frames with adjustable retention times, allowing selective retention and update of frames based on usage patterns, either in real-time (online) or offline, to optimize memory usage and compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If entire long-term reference frames are updated when scene changes occur, then compression efficiency is maintained, but processing cost and time consumption increase significantly
Solution Approach 1:
The patent segments the long-term reference frame into multiple versions (e.g., LTR frame 102, LTR frame 104, LTR frame 106) and selectively updates only the portions that have changed (such as updating LTR frame 104 to 106) rather than replacing the entire frame. This segmentation allows the system to maintain compression efficiency while reducing processing time by avoiding unnecessary updates of unchanged regions.
Solution Approach 2:
The patent applies local quality by updating only specific portions of the long-term reference frame that have changed (e.g., updating LTR frame 104 to 106) while retaining unchanged portions (LTR frame 102). This selective updating approach maintains compression efficiency for changed regions while avoiding the time cost of processing the entire frame.
2Productivity
If entire long-term reference frames are updated when scene changes occur, then compression efficiency is maintained, but data transmission cost increases
Solution Approach 1:
The patent extracts only the changed portions of the long-term reference frame (e.g., updating LTR frame 104 to 106) and transmits only these differences rather than the entire frame. This extraction approach maintains compression efficiency for the changed regions while significantly reducing data transmission cost by excluding unchanged portions.
Solution Approach 2:
The patent applies local quality by updating only specific portions of the long-term reference frame that have changed (e.g., updating LTR frame 104 to 106) while retaining unchanged portions (LTR frame 102). This selective updating approach maintains compression efficiency for changed regions while avoiding the time cost of processing the entire frame.
3Device complexity
If prior long-term reference frames are discarded when updated, then processing is simplified, but compression efficiency decreases when scenes repeat
Solution Approach 1:
The patent segments the long-term reference frame into multiple versions (e.g., LTR frame 102, LTR frame 104, LTR frame 106) and maintains all versions in the reference picture buffer. This segmentation allows the system to retain previous frames (LTR frame 102, LTR frame 104) even after updating (LTR frame 106), enabling efficient compression when scenes repeat without increasing processing complexity.
Solution Approach 2:
The patent performs preliminary action by maintaining multiple versions of long-term reference frames in the reference picture buffer before they are fully discarded. This allows the system to retain useful information from previous frames (LTR frame 102, LTR frame 104) for future reference, improving compression efficiency when scenes repeat without requiring complex processing.
4Productivity
If extended long-term reference frames with adjustable retention times are implemented, then compression efficiency and error-resilience are improved, but device complexity increases
Solution Approach 1:
The patent applies dynamics by implementing adjustable retention times for different long-term reference frames (e.g., LTR frame 102, LTR frame 104, LTR frame 106) based on their importance and usage patterns. This dynamic approach allows the system to adaptively manage reference frames, improving compression efficiency and error-resilience while controlling complexity through parameter adjustment rather than structural complexity.
Data Source
AI summary
An video signal processor includes circuitry configured to receive a video, determine a plurality of long-term reference frames, determine a long-term reference retention time, and encode the video into a bitstream including the determined plurality of long-term reference fames and the determined retention time. A decoded picture buffer for retaining a plurality of reconstructed pictures, including a long term reference picture, is also provided.