Subpixel 4-Tap Interpolation Filtering for Predictive Video Coding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video coding technologies face challenges in achieving efficient compression ratios with minimal sacrifice in picture quality, particularly in limited bandwidth scenarios, necessitating improved intra- and inter-prediction methods for video frames.

Innovation Solution

A method and apparatus for intra- or inter-prediction processing using a subpixel 4-tap interpolation filter with filter coefficients defined by analytical expressions, eliminating the need for look-up tables and reducing computational latency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional interpolation filters with fixed coefficients are used, then implementation is simple, but filtering accuracy varies across different subpixel positions

Engineering Contradiction:

Improvefiltering accuracyVSAvoidfilter coefficient calculation complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies dynamics by making the filter coefficients adaptive rather than fixed. The coefficients are dynamically calculated based on the subpixel offset value p, allowing the filter to adapt to different subpixel positions. This resolves the contradiction by enabling high filtering accuracy across all positions while maintaining a relatively simple implementation through the use of basic arithmetic operations.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of filter coefficients from fixed values to variable values that depend on the subpixel offset. By expressing coefficients as functions of p (e.g., c0 = (1+p)/4, c1 = (1-p)/4), the system achieves accurate filtering for any subpixel position without requiring complex lookup tables or iterative optimization.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If look-up tables are used to store filter coefficients for different subpixel positions, then filtering accuracy is maintained, but memory demands and access latency increase

Engineering Contradiction:

Improvefiltering accuracyVSAvoidmemory resources

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent extracts the essential information from look-up tables by deriving closed-form mathematical expressions for the filter coefficients. Instead of storing pre-computed coefficient values in memory, the system extracts the underlying pattern (coefficients as functions of p) and computes them on-demand using simple arithmetic operations, thereby eliminating the need for large coefficient tables.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent substitutes the mechanical system of look-up table storage and retrieval with a computational system based on mathematical formulas. Rather than physically storing and accessing coefficient tables in memory, the system uses algebraic expressions to generate coefficients, replacing memory-intensive operations with computation-efficient operations.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If complex filter coefficient calculations are performed, then filtering accuracy improves, but computational latency increases

Engineering Contradiction:

Improvefiltering accuracyVSAvoidcomputational latency

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by using only the necessary computational steps to achieve accurate filtering. The coefficient calculations use simple arithmetic operations (addition, subtraction, division by 4) that are sufficient for high accuracy without requiring complex computations. This avoids excessive computation while maintaining filtering precision.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

By changing the parameter representation to closed-form expressions of p, the system enables fast computation. The coefficients are calculated using direct arithmetic formulas rather than iterative methods or complex algorithms, significantly reducing computational latency while maintaining accuracy across all subpixel positions.

Inventive Principle:

Principle #35Parameter changes

4Productivity

If video data is heavily compressed to accommodate limited bandwidth, then transmission efficiency improves, but picture quality deteriorates

Engineering Contradiction:

Improvecompression ratioVSAvoidpicture quality

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent applies local quality by improving the prediction accuracy specifically at subpixel positions where interpolation is needed. By using adaptive filter coefficients that account for the fractional offset p, the system enhances the local quality of predicted samples without requiring overall increases in compression complexity or data rates.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS20250343954A1Method and apparatus of interpolation filtering for predictive coding

Publication Date: 2025.11.06 HUAWEI TECH CO LTD

AI summary

The present disclosure relates to intra- or inter-prediction for video encoding and decoding. For that purpose, an apparatus and methods obtain a reference sample and a subpixel offset value. A 4-tap interpolation filter is used to filter the reference sample to obtain a predicted sample value. The filter coefficients c0, c1, c2, and c3 of the 4-tap interpolation filter satisfyc0=1⁢6-p2,c1=1⁢6+1⁢6-p2,c2=1⁢6+p2,and⁢ c3=p2,with p being a fractional part of the value of the subpixel offset.