Affine Motion Prediction for Complex Image Block Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression technologies struggle to efficiently handle high-resolution and high-quality images, particularly in managing complex motion models such as rotation and zooming, which are difficult to process with traditional block-based prediction methods.
Innovation Solution
An affine model-based approach is employed to generate a candidate list for motion information prediction, derive control point vectors, and perform inter prediction using motion vectors, considering spatial and temporal candidates, and subblock units to enhance encoding/decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional block-based prediction methods are used, then the processing is simple, but the ability to handle complex motion models such as rotation and zooming is insufficient
Solution Approach 1:
The current block is divided into multiple subblocks, and motion compensation is performed separately for each subblock using affine motion vectors. This segmentation allows the system to handle complex motion models like rotation and zooming by applying different motion parameters to different regions, thereby improving adaptability while managing complexity through localized processing.
Solution Approach 2:
The patent introduces affine motion models that allow dynamic motion parameters (rotation, zooming, shearing) to be applied to different subblocks. Instead of using a single static motion vector for the entire block, the system dynamically adjusts motion parameters for each subblock based on local motion characteristics, enabling accurate representation of complex motion patterns.
2Measurement precision
If motion compensation is performed in units of subblocks, then the prediction accuracy is improved, but the computational complexity increases
Solution Approach 1:
The block is divided into subblocks for independent motion compensation processing. This segmentation improves prediction accuracy by capturing local motion variations within the block, while the computational complexity is managed by using efficient affine transformation calculations and reusing control point vectors across adjacent subblocks where possible.
Solution Approach 2:
Different motion parameters are applied to different subblocks based on their local motion characteristics. This local quality approach improves prediction accuracy by adapting motion models to local regions, while computational complexity is optimized by deriving control point vectors from neighboring blocks and using hierarchical processing strategies.
3Productivity
If an affine model is used for motion compensation, then the encoding/decoding performance is improved, but the processing complexity increases
Solution Approach 1:
The affine motion model introduces dynamic parameters (rotation, zooming, shearing) that improve encoding/decoding performance by accurately representing complex motion patterns. The processing complexity is managed by deriving control point vectors from neighboring blocks and using efficient affine transformation calculations that can be implemented with standard arithmetic operations.
Solution Approach 2:
The system derives control point vectors and motion parameters automatically from neighboring block information and temporal candidates without requiring manual intervention. This self-service approach improves encoding performance by adapting to local motion characteristics while managing complexity through automated parameter derivation and selection algorithms.
Data Source
AI summary
In an image encoding/decoding method and device according to the present invention, a candidate list for motion information prediction of a current block is generated, a control point vector of the current block is derived on the basis of the candidate list and a candidate index, a motion vector of the current block is derived on the basis of the control point vector of the current block, and inter-prediction with respect to the current block can be performed by means of the motion vector.