Motion Vector Difference Coding for Bi-Prediction Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality image/video data, particularly in immersive media formats like VR and AR, has led to higher transmission and storage costs due to the increased amount of information, and existing image/video compression technologies are inefficient in handling motion vector differences, especially in bi-prediction scenarios.
Innovation Solution
The method improves image/video compression efficiency by efficiently signaling motion vector differences and deriving L1 motion vector differences using short-term reference pictures, reducing coding system complexity through optimized inter prediction techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion estimation is performed on a block-by-block basis using traditional methods, then motion compensation can be achieved, but the computational complexity and processing time increase significantly
Solution Approach 1:
The image is divided into multiple blocks, and motion estimation is performed independently for each block. This segmentation allows parallel processing of different blocks, reducing overall computational complexity while maintaining motion compensation accuracy through block-level optimization.
Solution Approach 2:
Motion vectors for subsequent frames are predicted based on motion vectors from previous frames before actual motion estimation is performed. This preliminary action reduces the search space and computational requirements for full motion estimation, thereby reducing device complexity while maintaining reliability.
2Manufacturing precision
If high-resolution image processing is performed, then image quality is improved, but the processing time and computational resources required increase
Solution Approach 1:
Motion estimation is performed at a reduced resolution or with limited search ranges for certain blocks, applying partial action only where necessary. This reduces processing time while maintaining image quality by focusing computational resources on critical regions rather than uniformly processing the entire high-resolution image.
Solution Approach 2:
Motion vectors are predicted in advance using previous frame data before final image processing. This preliminary estimation reduces the computational burden during high-resolution processing, thereby reducing processing time while preserving image quality through subsequent refinement steps.
3Measurement precision
If motion vectors are calculated for every block in every frame, then motion accuracy is improved, but the data transmission requirements and bandwidth increase
Solution Approach 1:
Instead of transmitting complete motion vector data for every block, the system transmits differential values or simplified representations that can be reconstructed at the receiving end. This copying approach maintains motion accuracy through precise differential encoding while significantly reducing the quantity of data that needs to be transmitted.
Solution Approach 2:
Motion vectors for current frames are predicted based on previous frame data before transmission. Only the differences or corrections to these predicted vectors are transmitted, reducing data volume while maintaining motion accuracy through the preliminary prediction step that eliminates redundant information.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to embodiments of the present document, symmetric motion vector differences (SMVDs) and merge mode motion vector differences (MMVDs) according to an inter prediction mode may be performed on the basis of reference picture types and, particularly, short term reference pictures may be used. Therefore, prediction performance and coding efficiency in the inter prediction mode can be improved.