Image Coding Prediction With Third-Frame Boundary Pixel Interpolation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Motion compensation pixel interpolation in VVC cannot be applied when a scaling window is set to a reference picture and the reference picture is magnified or reduced, or when blocks inside the screen boundary of the reference picture do not include motion information, leading to reduced prediction accuracy in inter prediction.

Innovation Solution

An image encoding apparatus that generates a prediction image for a target block by referencing a second frame, interpolates pixels outside the boundary of the second frame using pixels from a third frame, and transforms the resolution of the frame to enhance interpolation accuracy and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If motion compensation pixel interpolation is used to generate pixels outside the screen boundary, then prediction accuracy is improved, but it cannot be applied when a scaling window is set or blocks lack motion information

Engineering Contradiction:
Improveprediction accuracyVSAvoidapplicability of interpolation method
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent changes the source of motion information from the current block to adjacent blocks when the current block lacks motion information. This parameter change in the data source allows the interpolation method to be applied universally across all blocks, resolving the contradiction between maintaining high prediction accuracy and ensuring broad applicability.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent introduces adjacent blocks as an intermediary source to provide motion information when the current block lacks it. This intermediary approach enables motion compensation pixel interpolation to function in cases where direct motion information is unavailable, thereby improving both prediction accuracy and method versatility.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of manufacture

If simple replication is used for pixels outside the screen boundary, then encoding is simplified, but prediction accuracy deteriorates

Engineering Contradiction:
Improveencoding simplicityVSAvoidprediction accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent changes the parameter of pixel generation from simple replication to motion compensation-based generation. By using motion information from blocks (either current or adjacent) to calculate pixel positions and values, the system achieves higher prediction accuracy while maintaining reasonable encoding complexity through efficient motion vector utilization.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If motion compensation pixel interpolation is applied, then encoding efficiency is improved, but it increases device complexity

Engineering Contradiction:
Improveencoding efficiencyVSAvoidinterpolation processing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the processing into distinct cases: blocks with motion information use direct motion compensation, while blocks without motion information use adjacent block motion information. This segmentation allows the system to apply the most efficient method for each case, improving overall encoding efficiency while managing device complexity through conditional processing logic.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4604536A1Image coding device, image coding method and program, image decoding device, and image decoding method and program
Publication Date: 2025.08.20 CANON KK
  • EP4604536A1 patent drawingFigure 1
  • EP4604536A1 patent drawingFigure 2
  • EP4604536A1 patent drawingFigure 3

AI summary

An image encoding apparatus comprises predicting means for generating a prediction image for a target block in a first frame to be encoded by referencing a second frame encoded before the first frame; encoding means for encoding a prediction error of the target block for the prediction image; interpolating means for, in a case where pixels outside of a boundary of the second frame are referenced, interpolating the pixels outside of the boundary of the second frame using pixels of a third frame encoded before the second frame; and transforming means for changing a resolution of a frame before the first frame.