Video Inter-Prediction With Bi-Directional Optical Flow
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression techniques, such as H.264/AVC and HEVC, face challenges in efficiently encoding and decoding high-resolution and high-frame-rate video data due to increasing data volumes, necessitating improved inter-prediction methods for better compression performance.
Innovation Solution
A coding tool that compensates for various motions of objects by using bi-directional optical flow and weighted bi-prediction, where the execution of bi-directional optical flow is conditioned on consistent luma and chroma weights across reference pictures, enhancing inter-prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If traditional inter-prediction methods are used, then encoding simplicity is maintained, but compression performance deteriorates for high-resolution and high-frame-rate video data
Solution Approach 1:
The patent segments the prediction process into multiple stages: generating initial prediction blocks from reference pictures, performing bidirectional optical flow analysis on these blocks, and combining results through weighted averaging. This segmentation allows complex motion compensation to be broken down into manageable steps that improve compression without overwhelming system resources.
Solution Approach 2:
The patent performs preliminary actions by generating prediction blocks from reference pictures before applying bidirectional optical flow analysis. This preliminary prediction provides a baseline that reduces the complexity of subsequent motion compensation steps, as the optical flow only needs to refine rather than create predictions from scratch.
2Measurement precision
If bidirectional optical flow is always applied, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by using weighted bi-prediction where different weights are assigned to different reference blocks based on their reliability. This allows the system to focus computational resources on areas where bidirectional optical flow provides the most benefit while using simpler prediction methods where they suffice, thereby improving accuracy without uniformly increasing complexity across the entire video stream.
Solution Approach 2:
The patent changes parameters dynamically by adjusting the weights applied to different reference pictures based on motion characteristics and prediction quality metrics. When motion is complex or reference quality is high, the system adjusts weights to optimize the contribution of each reference block, thereby adapting prediction accuracy to local conditions without always invoking full bidirectional optical flow computation.
3Loss of energy
If multiple reference pictures are used for bi-prediction, then compression performance is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by selectively applying bidirectional optical flow to only certain prediction blocks rather than uniformly processing all blocks. The system evaluates whether the additional processing time will yield sufficient compression improvement and applies the computationally intensive optical flow analysis only where necessary, thereby achieving better overall compression without proportionally increasing total processing time.
Data Source
AI summary
A video decoding apparatus, a video encoding apparatus, and an apparatus for transmitting a bitstream containing encoded video data are disclosed. The apparatuses are configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; generate a prediction block of the target block using the first and second reference blocks; generate a residual block based on the prediction block of the target block; and encode the residual block. The apparatuses are configured to generate the prediction block of the target block by performing a bi-directional optical flow process using the first and second reference blocks.