Texture Synthesis Polynomial Fitting for Complex Video Regions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video coding technologies, such as HEVC, struggle with efficiently encoding complex textures due to their predictability issues, leading to suboptimal coding efficiency and subjective quality in high-complexity and irregular textures.

Innovation Solution

Employing texture synthesis with luminosity adjustment and motion compensation using polynomial functions to approximate texture regions, transforming blocks into the spectral domain, and coding the parameters into a bitstream.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional hybrid video coding (HEVC) is used for encoding complex textures, then coding efficiency is improved for simple textures, but subjective quality deteriorates for high-complexity and irregular textures

Engineering Contradiction:
Improvecoding efficiencyVSAvoidsubjective quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent segments the video signal into two distinct types: complex textures and non-complex textures. This segmentation allows different encoding strategies to be applied to each type - texture synthesis for complex textures and conventional hybrid coding for non-complex textures, thereby resolving the contradiction between coding efficiency and subjective quality for different texture types

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the encoding parameters based on texture complexity. For complex textures, it uses texture synthesis parameters (patch coordinates, polynomial coefficients) instead of conventional prediction parameters, which fundamentally alters the encoding approach to achieve better subjective quality while maintaining acceptable coding efficiency

Inventive Principle:
Principle #35Parameter changes

2Manufacturing precision

If texture synthesis is used to improve subjective quality for complex textures, then perceived quality is enhanced, but bit rate increases due to parameter transmission

Engineering Contradiction:
Improvesubjective qualityVSAvoidbit rate
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent extracts only the essential parameters needed for texture synthesis (patch coordinates and polynomial coefficients) and transmits them in the bitstream, while the actual texture data is synthesized at the decoder side. This extraction approach reduces the bit rate compared to transmitting full texture data while maintaining subjective quality

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of transmitting the actual texture data and hoping for good reconstruction, the patent inverts the approach by transmitting synthesis parameters and generating the texture at the decoder. This inversion reduces bit rate while maintaining or improving subjective quality through accurate polynomial-based reconstruction

Inventive Principle:
Principle #13The other way round (Inversion)

3Device complexity

If polynomial fitting is used to model luminosity and motion in texture regions, then the number of parameters is reduced, but approximation accuracy may deteriorate for highly irregular textures

Engineering Contradiction:
Improvenumber of parametersVSAvoidapproximation accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent uses dynamic polynomial order selection, where the polynomial order is adapted to the local texture characteristics. For smoother regions, lower-order polynomials are used (reducing parameters), while for more irregular regions, higher-order polynomials are applied (improving accuracy), thus dynamically balancing parameter count and approximation accuracy

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies different polynomial orders to different regions of the texture based on local complexity. This local quality approach ensures that approximation accuracy is maintained in critical areas while using fewer parameters in simpler regions, resolving the contradiction between parameter reduction and accuracy preservation

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3662670B1Polynomial fitting for luminance reconstruction in texture synthesis
Publication Date: 2026.01.21 HUAWEI TECH CO LTD
  • EP3662670B1 patent drawingFigure 1~2
  • EP3662670B1 patent drawingFigure 3
  • EP3662670B1 patent drawingFigure 4

AI summary

The present disclosure relates to encoding a decoding video employing texture coding. In particular, a texture region is identified within a video picture and a texture patch is determined for said region. Moreover, a set of parameters specifies luminance within the texture region (1001) by fitting the texture region samples to a two-dimensional polynomial function of the patch determined according to the set of parameters (1040); and/or motion within the texture region by fitting motion estimated between the texture region of the video picture and an adjacent picture to a two-dimensional polynomial The texture patch and the first set of parameters are then included into a bitstream which is output of the encoder and provided in this way to the decoder which reconstructs the texture based on the patch and the function applied to the patch.