Video Codec Division Approximation Using Piecewise Inverse Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Division operations in video coding, such as those required for cross-component linear models in VVC/H.266 and ECM 3.0, are complex and resource-intensive, necessitating efficient approximations to reduce computational burden and memory requirements.
Innovation Solution
A method using piecewise approximations and look-up tables to derive scale and shift parameters for approximating division operations, allowing for division-free calculations through multiplications, additions, and bitwise shifts, thereby reducing the complexity and resource requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If division operations are used for cross-component linear models in video coding, then calculation accuracy is maintained, but computational complexity and resource requirements increase
Solution Approach 1:
The patent transforms the division operation parameters into multiplication parameters by pre-calculating inverse values. Instead of performing division during video coding, the system pre-computes 1/d values and stores them in lookup tables, changing the operational parameter from division to multiplication which is computationally simpler.
Solution Approach 2:
The patent performs the complex division operation in advance during a setup phase, calculating inverse values before the actual video coding process. This preliminary calculation allows the main coding process to use simple multiplication and lookup operations instead of complex division operations.
2Measurement precision
If division operations are used for cross-component linear models, then precision is maintained, but memory requirements increase
Solution Approach 1:
The patent segments the precision requirements into different lookup tables for different parameter types (alpha and beta parameters). Instead of storing all possible division results with full precision, the system creates separate lookup tables for different parameter ranges and precision levels, reducing overall memory requirements while maintaining necessary precision.
Solution Approach 2:
The patent uses lookup tables that store pre-computed inverse values as copies of calculated results. Instead of performing division operations and storing full-precision results, the system creates compact copies of inverse values in lookup tables that can be quickly retrieved and used multiple times, reducing memory usage while maintaining precision.
Data Source
AI summary
The embodiments relate to a method for encoding/decoding, the method comprising encoding/decoding (910) a picture comprising a number of samples, wherein a phase of the encoding/decoding comprises a division operation; determining (920) a numerator and a denominator; determining (930) an approximated output for the division operation between the numerator and the denominator, wherein the determining comprises deriving a scale parameter using a piecewise approximation; deriving a shift parameter and a rounding parameter, and applying the scale parameter, the shift parameter, and the rounding parameter to the numerator; using (940) the approximated output for the division operation in said phase of the encoding/decoding. The embodiments also relate to an apparatus and a computer program product for implementing the method.