Entropy Coding Probability Models for Time-Variant Video Streams
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital video compression techniques face challenges in efficiently reducing the data amount in video streams, particularly due to the time-variant nature of probability distributions in video data, which affects the accuracy of probability estimation for entropy coding.
Innovation Solution
A multimodal approach is described that uses multiple linear update models to accurately estimate probabilities for entropy coding, specifically by combining a first-order linear system with another model to form a higher-order linear system, enabling more accurate conditional probability modeling for streaming symbols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single probability model is used for entropy coding, then the device complexity is low, but the measurement precision of probability estimation deteriorates due to the time-variant nature of video data
Solution Approach 1:
The patent combines multiple probability models (first-order linear system and higher-order linear system) into a unified probability estimation framework. The encoder and decoder each maintain and synchronize multiple probability models, merging their capabilities to achieve more accurate time-variant probability estimation for entropy coding while managing computational complexity through coordinated updates.
2Measurement precision
If multiple probability models are used to improve probability estimation accuracy, then the measurement precision improves, but the use of energy increases due to additional computations
Solution Approach 1:
The patent implements dynamic probability model updating where the complexity of probability estimation adapts to the actual time-variant characteristics of the video data. The system dynamically adjusts between different probability models based on the observed data patterns, using more complex models when necessary for accuracy and simpler models when sufficient, thereby optimizing energy consumption while maintaining precision.
3Measurement precision
If a higher-order linear system is used for probability estimation, then the measurement precision of conditional probability modeling improves, but the device complexity increases
Solution Approach 1:
The patent segments the probability estimation process into distinct components: a first-order linear system for basic probability tracking and a higher-order linear system for capturing temporal dependencies. This segmentation allows each component to specialize in specific aspects of probability estimation, improving overall accuracy while enabling independent optimization and management of each model's complexity.
Data Source
AI summary
Entropy coding a sequence of symbols is described. A first probability model for entropy coding is selected. At least one symbol of the sequence is coded using a probability determined using the first probability model. The probability according to the first probability model is updated with an estimation of a second probability model to entropy code a subsequent symbol. The combination may be a fixed or adaptive combination.