Joint Reference Picture Reordering Using Bilateral Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently utilizing reference pictures for inter-prediction, leading to suboptimal compression and quality degradation due to inefficient use of motion vectors in video encoding and decoding processes.
Innovation Solution
Implementing a video processing method that generates a joint list of reference pictures based on bilateral matching using motion vectors, reorders these pairs to optimize inter-prediction, and determines a reference index for improved video block prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If reference pictures are used in traditional order for inter-prediction, then the encoding process is simple, but video compression efficiency is suboptimal
Solution Approach 1:
The patent applies dynamics by making the reference picture list order adaptive rather than fixed. The joint list dynamically reorders reference pictures based on bilateral matching costs calculated from motion vectors, allowing the system to adapt to different video content characteristics and optimize compression efficiency for each block
Solution Approach 2:
The patent changes the ordering parameter of reference pictures based on bilateral matching costs. By calculating matching costs between current block and reference blocks using motion vectors, the system reorders reference pictures to prioritize those with better matching characteristics, improving prediction accuracy
2Measurement precision
If motion vectors are used efficiently with reordered reference pictures, then inter-prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary action by calculating bilateral matching costs and reordering reference pictures before the actual inter-prediction process. This pre-ordering ensures that when motion vectors are applied, they reference the most suitable pictures, improving accuracy without adding complexity during the main prediction step
Solution Approach 2:
The patent introduces bilateral matching cost as an intermediary parameter that mediates between motion vectors and reference picture selection. This intermediary enables more accurate matching by providing a quantitative measure of compatibility between current and reference blocks
3Adaptability or versatility
If joint list with multiple reference picture pairs is generated, then prediction flexibility increases, but bit rate increases
Solution Approach 1:
The patent applies partial action by generating a joint list with multiple reference picture pairs but only signaling the necessary information for the selected pair. The system creates comprehensive candidate lists for flexibility but transmits only minimal data (reference picture indices) for actual prediction, avoiding excessive bit rate increase
Data Source
AI summary
Implementations of the disclosure provide video processing apparatuses and methods for inter-prediction. A video processing method for the inter-prediction performed by a video decoder includes determining a reference index for a video block in a video frame of a video from a bitstream. The video processing method also includes generating a joint list including a plurality pairs of reference pictures based on a first reference list and a second reference list for the video block. The video processing method further includes reordering the plurality pairs of reference pictures in the joint list based on bilateral matching using one or more motion vectors of the video block to generate a reordered joint list. The video processing method further includes performing the inter-prediction on the video block based on the reordered joint list and the reference index.