Buffered Picture Memory Reduction via Region-Based Down-Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multiple reference frame architectures in motion compensation face challenges with high memory access rates and insufficient memory bandwidth, leading to degraded performance and unwanted side effects when trying to reduce memory requirements, such as picture quality degradation and increased computation load.
Innovation Solution
A method and apparatus that decodes header information to identify less important portions of moving pictures, buffers them, and performs post-processing operations to reduce memory requirements, using look-ahead information to determine which data to discard or compress, thereby optimizing memory usage without degrading picture quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If decoded pictures are scaled down to reduce memory requirement, then memory bandwidth consumption is reduced, but picture quality is degraded
Solution Approach 1:
The patent divides the decoded picture into different importance regions (important region and less important region) based on motion vector analysis. Only the less important region undergoes down-sampling processing, while the important region maintains full resolution. This segmentation approach reduces overall memory bandwidth consumption while preserving picture quality in critical areas.
Solution Approach 2:
Different processing strategies are applied to different regions of the picture based on their importance. The important region (containing significant motion or detail) is preserved at full resolution, while the less important region is down-sampled. This local quality differentiation allows memory reduction without uniformly degrading picture quality.
2Quantity of substance
If decoded pictures are compressed in a simpler way to reduce memory requirement, then memory access rate is reduced, but picture quality is degraded
Solution Approach 1:
The patent segments the picture into important and less important regions, applying different processing to each. The less important region is subjected to selective down-sampling while the important region maintains original quality, thus reducing memory requirements without uniformly compromising picture quality.
Solution Approach 2:
The patent changes the resolution parameter selectively for different regions of the picture. The less important region undergoes resolution reduction (down-sampling) while the important region maintains its original resolution parameters, achieving memory optimization without significant quality loss.
3Quantity of substance
If just-in-time decoding of specific frames is applied to reduce memory requirement, then memory bandwidth is reduced, but computation load becomes extremely heavy
Solution Approach 1:
The patent performs preliminary analysis of motion vectors during the decoding process to identify important and less important regions before applying down-sampling. This preliminary action allows the system to make informed decisions about which regions can be compressed, reducing computation load compared to exhaustive just-in-time decoding approaches.
Solution Approach 2:
The patent dynamically changes processing parameters (resolution levels) based on motion vector analysis results. By adjusting the down-sampling ratio for different regions based on their motion characteristics, the system reduces computation load while maintaining picture quality where needed.
Data Source
AI summary
A method for storing decoded moving pictures with a reduced memory requirement includes: decoding header information in a bitstream; storing a decoded moving picture as a buffered picture; determining a less important portion of the buffered picture; and performing a post-processing operation on the less important portion of the buffered picture according to the header information. An associated apparatus for storing decoded moving pictures with a reduced memory requirement is further provided.