Motion Vector Prediction for Non-Uniform Video Motion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards are suboptimal for handling non-uniform motion fields, particularly those caused by camera motion, leading to inefficient compression and coding in video frames with translational or rotational movements.
Innovation Solution
A method is introduced that selects between different motion vector predictor algorithms based on the presence of non-uniform motion in a video frame, using either a parametric model or spatial/temporal candidates to improve prediction accuracy and efficiency, especially for spherical video projections like equirectangular formats.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional motion vector prediction algorithms are used, then uniform translational motion can be effectively handled, but non-uniform motion fields caused by camera motion result in suboptimal coding efficiency
Solution Approach 1:
The patent applies dynamics by making the motion vector prediction algorithm adaptive and changeable based on the detected motion characteristics. The system dynamically selects between different prediction algorithms (spatial, temporal, or combined) depending on whether the motion field is uniform or non-uniform, allowing the coding system to adapt to varying motion patterns in different video sequences and blocks.
Solution Approach 2:
The patent changes the parameter of motion vector prediction by introducing motion field analysis that detects uniformity characteristics. Based on this detection, the system changes the prediction algorithm parameters - using spatial prediction for uniform motion, temporal prediction for non-uniform motion, or combined approaches intermediate cases - thereby optimizing coding efficiency for different motion scenarios.
2Loss of information
If motion vector compression schemes use already encoded blocks as references, then spatial correlation can be exploited for uniform motion, but accuracy deteriorates for non-uniform motion fields
Solution Approach 1:
The patent introduces motion field analysis as an intermediary that assesses the uniformity of motion characteristics before selecting the prediction approach. This intermediary analysis enables the system to choose appropriate reference blocks and prediction algorithms, thereby maintaining both compression efficiency and prediction accuracy across different motion types.
Solution Approach 2:
The patent segments the video content into blocks with different motion characteristics by analyzing motion field uniformity. Each segment is then processed with an appropriate prediction algorithm - spatial prediction for uniform motion segments, temporal prediction for non-uniform segments - ensuring both compression efficiency and accuracy are optimized for each segment's specific characteristics.
3Device complexity
If a single motion vector prediction algorithm is used, then implementation is simple, but coding performance suffers for videos with complex camera movements
Solution Approach 1:
The patent implements a dynamic algorithm selection mechanism that chooses between spatial, temporal, or combined prediction algorithms based on real-time analysis of motion field uniformity. This dynamic approach maintains reasonable implementation complexity by using a decision-based framework rather than simultaneously implementing multiple complex algorithms, while significantly improving compression efficiency for videos with complex camera movements.
Solution Approach 2:
The patent performs preliminary analysis of motion field uniformity before selecting the prediction algorithm. This preliminary action allows the system to pre-determine the most appropriate prediction approach for each block, avoiding the need for complex real-time adjustments during encoding and maintaining implementation simplicity while achieving high compression efficiency.
Data Source
AI summary
Methods and systems for providing a bitstream comprising video data encoded by an encoder apparatus are described, wherein the method may include: a processor of the encoder apparatus determining a current motion vector of a current block of a current video frame of a sequence of video frames comprising video data, the current motion vector defining a spatial offset of the current block relative to a prediction block of a previously encoded reference video frame stored in a memory of the encoder apparatus; the processor determining or receiving motion information associated with the current video frame, the motion information signaling the processor whether at least part of the offset defined by the current motion vector is associated with non-uniform motion in the video data of the current video frame; the processor determining a motion vector predictor candidate based on the motion information, at least a first motion vector predicator algorithm and a second motion vector predictor algorithm; and, the processor determining a motion vector difference based on the selected motion vector predictor candidate and the current motion vector; and, the processor using an encoding process to encode a residual block, the motion vector difference, an indication of the selected motion vector predictor candidate, and at least part of the motion information into a bitstream, wherein the residual block defines a difference between the current block and the prediction block.