Hierarchical Image Coding Units for Quality and Overhead Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image decoding and coding methods struggle to simultaneously improve image quality and coding efficiency, as parameters like quantization scale parameters are applied uniformly across large units, preventing adjustment for smaller processing units and increasing coding overhead.
Innovation Solution
An image decoding and coding method that hierarchically layers processing units, allowing parameters to be stored and applied differently at each layer, with hierarchy depth information specifying the layer, enabling flexible parameter application and reducing coding overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If parameters are applied uniformly across large coding units to improve coding efficiency, then coding overhead is reduced, but image quality cannot be optimized for smaller processing units
Solution Approach 1:
The patent segments the picture into multiple coding units of different sizes (e.g., 64x64, 32x32, 16x16 macroblocks) and stores quantization parameters at different hierarchical levels. This allows fine-grained control of image quality for small units while maintaining efficient coding for large units, resolving the contradiction between image quality optimization and parameter storage complexity.
Solution Approach 2:
The patent introduces a hierarchical dimension for parameter storage, organizing quantization parameters across multiple levels (first level for large units, second level for smaller units). This dimensional approach enables selective parameter application at appropriate scales, improving image quality without proportionally increasing overall parameter storage complexity.
2Manufacturing precision
If parameters are stored for each processing unit to enable fine-grained quality control, then image quality improves, but coding overhead increases
Solution Approach 1:
The patent applies partial parameter storage by storing quantization parameters only at hierarchical levels where they provide meaningful control. Not all processing units require separate parameter storage, allowing fine-grained quality control where needed while avoiding redundant parameters where uniform treatment is sufficient, thus reducing overall coding overhead.
Solution Approach 2:
The patent implements local quality control by storing different quantization parameters for different coding units at appropriate hierarchical levels. This allows optimization of image quality in specific regions or units without requiring parameters for all units, reducing unnecessary coding overhead while maintaining quality where it matters.
3Productivity
If large coding units are used to improve coding efficiency, then coding overhead is reduced, but flexibility to adjust parameters for smaller units is lost
Solution Approach 1:
The patent introduces dynamic parameter selection where the quantization parameter applied to a coding unit can be selected from different hierarchical levels based on the unit size and requirements. This dynamic approach maintains coding efficiency for large units while providing flexibility to adjust parameters for smaller units when needed, resolving the contradiction between coding efficiency and adaptability.
Data Source
AI summary
An image decoding method which can improve both image quality and coding efficiency is an image decoding method for decoding a coded stream which includes a plurality of processing units and a header for the processing units, the coded stream being generated by coding a moving picture, the processing units including at least one processing unit layered to be split into a plurality of smaller processing units, the image decoding method including specifying a hierarchical layer having a processing unit in which a parameter necessary for decoding is stored, by parsing hierarchy depth information stored in the header, and decoding the processing unit using the parameter stored in the processing unit located at the specified hierarchical layer.