Locally Adaptive Warped Motion Compensation for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression schemes are inefficient in handling non-translational motion within video frames, as they assume purely translational motion, leading to increased data requirements and processing complexity.
Innovation Solution
The implementation of locally-adaptive warped motion compensation, where projection samples are used to determine a warping model for neighbor blocks to generate a prediction block by projecting pixels of a current block to a warped patch within a reference frame, allowing for non-rectangular patches and improved motion prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional translational motion compensation is used, then the encoding and decoding process is simple, but the data requirements increase and compression efficiency decreases when handling non-translational motion
Solution Approach 1:
The patent applies local quality by using different motion compensation models for different regions. Specifically, it identifies regions with non-translational motion and applies warped motion compensation only to those regions, while using traditional translational motion compensation for other regions. This localized approach improves compression efficiency for complex motion areas without unnecessarily increasing complexity across the entire video frame.
Solution Approach 2:
The patent changes the motion model parameters from simple translational vectors to warped projection parameters. By introducing warping parameters that describe non-linear motion transformations, the system can accurately represent complex motion patterns. The encoder determines warped motion parameters for blocks exhibiting non-translational motion, and the decoder uses these parameters to reconstruct the warped regions, thereby improving compression efficiency.
2Measurement precision
If warped motion compensation is applied to all blocks, then motion prediction accuracy improves, but processing complexity increases significantly
Solution Approach 1:
The patent applies partial action by using warped motion compensation only for blocks that exhibit non-translational motion characteristics, rather than applying it universally. The encoder evaluates motion characteristics of each block and selectively applies the warped model only where needed, achieving high motion prediction accuracy for complex regions while avoiding unnecessary processing complexity for simple translational regions.
Solution Approach 2:
The patent segments the video frame into different motion regions and applies appropriate compensation models to each segment. Blocks are divided into those requiring warped motion compensation and those suitable for traditional translational compensation. This segmentation allows the system to achieve high overall prediction accuracy by matching the right model to each region without uniformly increasing processing complexity across the entire frame.
3Manufacturing precision
If non-rectangular warped patches are used, then the accuracy of predicting non-translational motion improves, but the device complexity and computational resources increase
Solution Approach 1:
The patent introduces dynamic adaptation by allowing the motion compensation model to change its complexity based on the actual motion characteristics of each block. The system dynamically selects between rectangular and non-rectangular warped patches, and between different warping model complexities, based on the measured motion patterns. This dynamic approach achieves high modeling precision for complex motions while maintaining lower complexity for simpler cases.
Solution Approach 2:
The patent changes geometric parameters of the compensation regions from fixed rectangular shapes to flexible non-rectangular warped patches. By adjusting the shape parameters to match the actual motion deformation, the system achieves higher motion modeling precision. The encoder determines appropriate warping parameters including non-rectangular geometry when needed, and the decoder reconstructs these varied shapes, improving precision without permanently increasing device complexity.
Data Source
AI summary
Encoding or decoding blocks of video frames using locally adaptive warped motion compensation can include determining projection samples for predicting a warped motion of a current block to be encoded or decoded based on a warping model of a neighbor block adjacent to the current block. Parameters of a projection model can be determined based on the projection samples. A prediction block can be generated by projecting pixels of the current block to a warped patch within a reference frame using the parameters of the projection model. The warped patch can be a non-rectangular patch having a shape and a position in the reference frame indicated by the parameters of the projection model.