Video Decoding Sign Prediction for Lower Residual Derivation Cycles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The exponential increase in the number of signs to be predicted in existing video coding schemes leads to a significant increase in calculation amount and processing cycles due to sign prediction being added to quantization and transform processing.
Innovation Solution
A video decoding apparatus that includes a parameter decoder to decode transform coefficients and a sign prediction unit to predict the sign of the coefficients, utilizing parameters such as quantization parameters, last positions, and syntax elements to optimize sign prediction, thereby reducing the number of signs to be predicted.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of signs to be predicted increases to improve prediction accuracy, then the prediction precision is improved, but the amount of calculation increases exponentially
Solution Approach 1:
The patent segments the sign prediction process by dividing transform coefficients into different groups based on their positions (e.g., high-frequency coefficients, low-frequency coefficients, boundary coefficients). Different prediction methods or levels of prediction are applied to different segments, thereby reducing the overall computational complexity while maintaining prediction accuracy for the most important coefficients.
Solution Approach 2:
The patent applies local quality by using different prediction strategies for different regions or types of coefficients. For example, coefficients near block boundaries may use one prediction method while interior coefficients use another, or high-frequency coefficients may be predicted differently from low-frequency ones. This localized approach optimizes the balance between precision and computational load.
2Productivity
If sign prediction is added to quantization and transform processing to improve coding efficiency, then the coding efficiency is improved, but the number of cycles required to derive residuals increases
Solution Approach 1:
The patent performs preliminary actions by preparing prediction data and structures in advance, such as pre-computing boundary pixel values or organizing coefficient data in a manner that facilitates faster prediction during the main decoding process. This preliminary preparation reduces the computational burden during residual derivation.
Solution Approach 2:
The patent implements skipping mechanisms where certain prediction steps are omitted or simplified under specific conditions. For example, if a coefficient is known to have a predictable sign based on neighboring coefficients, the full prediction computation may be skipped. This allows the system to rush through less critical computations while maintaining accuracy where needed.
Data Source
Figure 1
Figure 2(a)~2(f)
Figure 3
AI summary
Video coding and decoding apparatuses that can enhance coding efficiency are provided. A video decoding apparatus includes a parameter decoder configured to decode a transform coefficient and a sign prediction parameter from coded data, and a sign prediction unit configured to predict a sign of the transform coefficient. For a candidate of a combination of signs of transform coefficients without sign information, the sign prediction unit derives a cost with reference to a boundary pixel with a neighboring block, predicts the sign based on the cost, and modifies the predicted sign of the transform coefficient using the sign prediction parameter obtained by decoding by the parameter decoder.