Decoding Reference Picture Buffering for Video Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital video compressed with block-based techniques exhibits blocking artifacts, particularly noticeable in codecs without normative deblocking filters, leading to unacceptable drift and degraded video quality when trying to reconstruct blocky frames from non-deblocked reference frames.
Innovation Solution
A method involving the generation and buffering of both non-post-processed and post-processed reference pictures during decoding, allowing for the selection of the appropriate reference type based on picture type, to reduce blockiness and prevent drift accumulation, thereby improving visual quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If in-loop deblocking is applied to all frames during decoding, then blockiness is reduced for codecs with normative in-loop deblocking, but drift error accumulates until the next refresh picture
Solution Approach 1:
The patent segments the reference picture storage into two distinct buffers: a first buffer storing non-deblocked reference pictures and a second buffer storing deblocked reference pictures. This segmentation allows the system to selectively access appropriate reference pictures based on picture type, preventing drift accumulation while maintaining blockiness reduction benefits.
Solution Approach 2:
The patent applies different quality processing to different parts of the reference picture sequence. Non-reference pictures (B-pictures) are decoded using deblocked reference pictures to reduce blockiness, while reference pictures (P and I pictures) are decoded using non-deblocked reference pictures to prevent drift accumulation. This local quality approach optimizes each picture type's decoding quality according to its specific requirements.
2Manufacturing precision
If conventional post-processing deblocking is used, then blockiness in decoded pictures is reduced, but blocky reference frames are used to reconstruct B-frames causing interior blockiness
Solution Approach 1:
The patent performs deblocking processing on reference pictures before they are used for motion-compensated prediction of B-frames. By preliminarily removing blocking artifacts from reference pictures, the system prevents these artifacts from being propagated into the interior regions of B-frames during motion compensation, thereby eliminating interior blockiness.
Solution Approach 2:
The patent changes the state of reference pictures by applying deblocking filtering, transforming them from a blocked state to a smoothed state. This parameter change in the reference picture quality directly improves the quality of reconstructed B-frames by preventing the transmission of blockiness through the motion compensation process.
3Quantity of substance
If only a single version of each reference picture is generated, then memory usage is reduced, but both drift accumulation and interior blockiness cannot be simultaneously avoided
Solution Approach 1:
The patent adds a temporal dimension to reference picture storage by maintaining two distinct versions of reference pictures at different processing stages (deblocked and non-deblocked). This dimensional expansion allows the system to access the appropriate version based on the decoding requirements of different picture types, simultaneously avoiding drift accumulation and interior blockiness.
Solution Approach 2:
The patent creates two versions of reference pictures with different processing parameters: one version remains in its original non-deblocked state for reference picture decoding, while the other version undergoes deblocking processing for non-reference picture decoding. This parameter differentiation allows optimal quality for each picture type without compromising memory efficiency excessively.
Data Source
AI summary
A method of compressed picture reconstruction using a plurality of post-processed reference pictures. The method generally includes the steps of (A) generating a first of a plurality of reconstructed pictures by decoding a first of a plurality of compressed pictures using at least one of a plurality of non-post-processed reference pictures buffered in a reference memory, wherein the compressed pictures are received in an input bitstream, (B) generating a first of a plurality of processed pictures by artifact processing the first reconstructed picture to remove artifacts, (C) buffering in the reference memory both (i) the first reconstructed picture as one of the non-post-processed reference pictures and (ii) the first processed picture as one of the post-processed reference pictures and (D) generating a second of the reconstructed pictures by decoding a second of the compressed pictures using at least one of the post-processed reference pictures buffered in the reference memory.