Bidirectional Optical Flow Estimation Subblock Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The bidirectional optical flow (BIO) estimation method in video encoding is computationally complex, which increases costs and processing requirements.
Innovation Solution
A method for encoding or decoding video data using bidirectional optical flow (BIO) that determines motion vectors for subblocks or pixels within a current block, generating prediction blocks by minimizing the sum of squares of flow differences within a masking window, thereby reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If bidirectional optical flow (BIO) estimation is applied for motion compensation, then motion prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent divides the current block into multiple subblocks and performs BIO estimation separately for each subblock. This segmentation reduces the computational complexity by breaking down the large-scale optimization problem into smaller, independent subproblems, while still maintaining accurate motion prediction at the pixel level within each subblock.
Solution Approach 2:
The patent applies different BIO estimation processes to different subblocks within the current block, allowing each subblock to have its own motion characteristics. This local approach improves motion prediction accuracy for regions with varying motion patterns while reducing overall computational complexity by not requiring uniform high-precision estimation across the entire block.
2Manufacturing precision
If BIO process is applied in pixel unit, then motion compensation precision is improved, but processing complexity increases
Solution Approach 1:
The patent performs BIO estimation at the subblock level rather than for every individual pixel, which reduces processing complexity. The subblock-level estimation provides sufficient motion compensation precision for practical video encoding while avoiding the excessive computational burden of pixel-level BIO for all pixels in the block.
Solution Approach 2:
The patent applies BIO estimation to selected subblocks within the current block rather than uniformly to all blocks. This partial application approach maintains motion compensation precision where it is most needed while reducing overall processing complexity by skipping BIO estimation in regions where simpler methods suffice.
Data Source
AI summary
Disclosed herein is a method for adaptive bidirectional optical flow estimation for inter prediction compensation during video encoding. The method aims to reduce complexity and/or cost of bidirectional optical flow (BIO) at a pixel level or a subblock level.