Local Warp Models for Complex-Motion Video Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding techniques struggle to accurately capture the motion of blocks with complex motion patterns, particularly in video sequences with large blocks or low bit rates, leading to inefficient inter-prediction.
Innovation Solution
The use of a local warp model is generated by extending the motion model of an adjacent block to predict the current block, incorporating translational motion vectors and additional transformations like rotation, scaling, and shearing, ensuring continuity of motion across block boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional translational motion models are used for inter-prediction, then the encoding process is simple and computationally efficient, but the prediction accuracy for blocks with complex motion patterns deteriorates
Solution Approach 1:
The patent divides the motion model into two components: a translational motion vector (simple component) and a warp model (complex component). The translational part handles basic motion, while the warp model handles complex deformations. This segmentation allows the system to achieve high prediction accuracy for complex motion patterns while keeping the overall structure manageable by separating simple and complex elements.
Solution Approach 2:
The patent applies local quality by using a warp model specifically for blocks that exhibit complex motion patterns, rather than applying complex models universally. The warp model parameters are determined locally for each block based on its motion characteristics, allowing high prediction accuracy where needed while maintaining simplicity for blocks with straightforward motion.
2Measurement precision
If warp models with multiple parameters are used to capture complex motion, then prediction accuracy improves, but the number of bits required for signaling increases
Solution Approach 1:
The patent uses partial action by selectively applying the warp model only when complex motion is detected, rather than applying it to all blocks. The translational motion vector handles the bulk of motion compensation, and the warp model is used partially for blocks requiring additional accuracy. This reduces the average bitrate increase while maintaining high prediction accuracy where needed.
Solution Approach 2:
The patent changes parameters by using a simplified representation of the warp model where parameters are derived from the translational motion vector and adjacent block information. Instead of encoding full six-parameter warp models independently, the system derives parameters through mathematical relationships, reducing the number of bits required for signaling while maintaining prediction accuracy.
3Reliability
If motion continuity across block boundaries is enforced, then visual quality at boundaries improves, but the computational complexity of generating the warp model increases
Solution Approach 1:
The patent applies preliminary action by first determining the translational motion vector and the motion model of adjacent blocks before generating the current block's warp model. This preliminary information is used to constrain and guide the warp model generation process, ensuring boundary continuity is built-in from the start rather than requiring complex post-processing adjustments.
Solution Approach 2:
The patent uses the translational motion vector as an intermediary between the simple motion compensation and the complex warp model. The translational vector serves as a bridge that provides a base motion estimate, which then informs the warp model parameters to ensure continuity at block boundaries. This intermediary simplifies the generation process by providing a starting point for the warp model.
Data Source
AI summary
A warp model is determined for a block to be coded by using local motion extension. Motion for an adjacent block is extended locally into the block to determine motion for predicting the block. A motion model used for inter-prediction of an adjacent block that shares at least a portion of a boundary with a current block to be predicted is determined. The motion model is a first translational motion vector or a first warp model. A second translational motion vector for inter-prediction of the current block is determined. A second warp model for the inter-prediction of the current block is determined using the motion model and the second translational motion vector. An inter-prediction process for the current block is performed using the second warp model. For example. the interprediction process may include determining a prediction block for encoding or decoding the current block.