Image Coding Transform Component Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image coding schemes face inefficiencies in representing and coding motion information, particularly when capturing images with scaling or rotational motions, as they rely solely on translation motion, leading to increased data and processing requirements when using high-order motion information like affine transforms.
Innovation Solution
An image coding method that selects and codes specific transform components such as translation, rotation, scaling, and shearing components, allowing for improved prediction accuracy and reduced data by dividing motion information into these components and selectively using them based on block size and priority, thereby reducing the amount of information and processing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-order motion information like affine transforms is used to represent scaling or rotational motions, then prediction accuracy is improved, but data amount and processing requirements increase
Solution Approach 1:
The motion information is segmented into multiple transform components (translation, rotation, scaling, shearing) that can be selectively coded. Instead of using a complete affine transform for all blocks, the patent divides the motion representation into independent components that can be individually selected and coded based on the actual motion characteristics of each block, thereby reducing the overall data amount while maintaining prediction accuracy where needed
Solution Approach 2:
Different transform components are applied to different blocks based on their local motion characteristics. The patent selectively uses rotation, scaling, or shearing components only for blocks that exhibit corresponding motion patterns, while using simpler translation for other blocks. This local adaptation maintains high prediction accuracy for complex motions while reducing data requirements for simpler regions
2Measurement precision
If high-order motion information like affine transforms is used to represent scaling or rotational motions, then prediction accuracy is improved, but processing load increases
Solution Approach 1:
The processing of motion information is segmented into separate transform components that can be independently selected and processed. This segmentation allows the decoder to process only the necessary components for each block based on selection information, reducing the overall processing load compared to always processing complete affine transforms for all blocks
Solution Approach 2:
The patent applies partial action by selectively using only the necessary transform components for each block based on its motion characteristics. Instead of applying all affine transform operations to every block, the system performs only the required transformations (translation, rotation, scaling, or shearing) based on coded selection information, thereby reducing processing load while maintaining prediction accuracy
3Measurement precision
If multiple transform components are coded for each block, then prediction accuracy is improved, but coding complexity increases
Solution Approach 1:
The coding process is segmented into selecting and coding only the necessary transform components for each block based on its motion characteristics. The patent uses selection information to indicate which transform components (translation, rotation, scaling, shearing) are present, allowing the encoder to code only those components rather than always coding complete affine transforms, thereby reducing coding complexity while maintaining prediction accuracy
Solution Approach 2:
The patent applies partial action by coding only the necessary transform components for each block based on its actual motion characteristics. Instead of coding all possible transform parameters for every block, the system codes only the required components (e.g., only rotation for rotational motion, only scaling for zooming), reducing coding complexity while maintaining prediction accuracy through selective coding
Data Source
AI summary
An image coding method includes selecting two or more transform components from among a plurality of transform components that include a translation component and non-translation components, the two or more transform components serving as reference information that represents a reference destination of a current block; coding selection information that identifies the two or more transform components that have been selected from among the plurality of transform components; and coding the reference information of the current block by using reference information of a coded block different from the current block.