Unified Significance Map Coding With Shared Context Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards, such as MPEG-4 AVC and KTA, face challenges in effectively capturing the difference in coefficient distributions across various transform sizes during significance map coding, leading to inefficient context modeling and difficulty in extending designs to future standards with more transforms.
Innovation Solution
A unified rule-based approach is proposed for generating context sharing maps that adapt to different transform sizes, allowing multiple transform coefficient positions to share contexts, simplifying the design of encoders and decoders, and enabling easy extension to future standards.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate context models are used for different transform sizes, then the accuracy of capturing coefficient distributions is improved, but the device complexity and number of context models increases
Solution Approach 1:
The patent merges context modeling across different transform sizes by introducing a transform size indicator that allows the same context model to be applied to multiple transform sizes. Instead of maintaining separate context models for each transform size (4x4, 8x8, 16x16, 32x32), the system combines them into unified context models that are selected based on the transform size indicator, thereby reducing the total number of context models while maintaining adaptive modeling capability.
Solution Approach 2:
The patent makes context models universal across different transform sizes by designing context models that can serve multiple transform sizes simultaneously. The context models are no longer dedicated to specific transform sizes but are instead configured to work with multiple transform sizes through the transform size indicator mechanism, allowing one context model to perform multiple functions across different transform configurations.
2Measurement precision
If separate context models are used for different transform sizes, then the accuracy of capturing coefficient distributions is improved, but the ease of extending to future standards decreases
Solution Approach 1:
The patent creates a universal context modeling framework that can accommodate multiple transform sizes through the transform size indicator. This universal design allows future standards to easily extend to new transform sizes without requiring completely new context models, as the existing framework can be configured to handle new transform sizes by simply updating the transform size indicator mappings.
Solution Approach 2:
The patent introduces dynamic adaptability by allowing the context model selection to be configured based on the transform size indicator. This dynamic configuration enables the system to adapt to future transform sizes without structural changes to the core context modeling mechanism, making the system extensible and flexible for future standard evolution.
3Device complexity
If limited context templates are used, then the model cost is reduced, but the accuracy of probability estimation decreases
Solution Approach 1:
The patent changes the parameter of context model configuration by introducing transform size as a configurable parameter that influences context model selection. Instead of using a fixed limited set of context templates, the system dynamically selects or configures context models based on the transform size indicator, allowing the model to adapt its complexity and granularity to match the specific transform size being processed, thereby maintaining accuracy while controlling overall model cost.
Data Source
AI summary
Methods and apparatus are provided for unified significance map coding. An apparatus includes a video encoder (400) for encoding transform coefficients for at least a portion of a picture. The transform coefficients are obtained using a plurality of transforms. One or more context sharing maps are generated for the transform coefficients based on a unified rule. The one or more context sharing maps are for providing at least one context that is shared among at least some of the transform coefficients obtained from at least two different ones of the plurality of transforms.