Frame Decoding with Confidence-Weighted Prediction Residuals

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image compression technologies face poor compression performance due to inaccuracies in predicted values for frames with irregular motion or blocked areas, leading to significant coding errors.

Innovation Solution

The proposed method involves configuring a confidence value for predicted values, using a decoder-side apparatus with modules for entropy decoding, motion compensation, and processing to determine reconstructed frames based on predicted values, residuals, and confidence, thereby reducing coding errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of substance

If motion estimation and motion compensation technologies are used to predict current frame, then data redundancy is reduced and transmission costs are lowered, but coding error increases for areas with irregular motion or blocked areas

Engineering Contradiction:
Improvedata redundancyVSAvoidprediction accuracy
Core Design Contradiction:
Loss of substanceVSReliability

Solution Approach 1:

The patent introduces a confidence map as an intermediary component that mediates between the predicted frame and the residual coding process. The confidence map provides reliability information for each pixel in the predicted frame, allowing the decoder to appropriately handle areas with low prediction confidence (such as irregular motion or blocked areas) versus high confidence areas, thus resolving the contradiction between compression efficiency and prediction accuracy

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the parameter space by introducing confidence values as an additional parameter alongside the predicted frame and residual. This allows the system to dynamically adjust the weighting between predicted values and residual values based on the confidence map, enabling adaptive handling of different regions with varying prediction quality

Inventive Principle:
Principle #35Parameter changes

2Productivity

If residual coding is used to encode predicted error, then compression performance is improved for accurate predictions, but coding error becomes significant when predicted value is inaccurate or unavailable

Engineering Contradiction:
Improvecompression performanceVSAvoidcoding error
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent makes the reconstruction process dynamic by introducing confidence-based weighting. Instead of a static approach where residual coding is always applied uniformly, the system dynamically adjusts the contribution of predicted values versus residual values based on the confidence map, allowing adaptive optimization of compression performance and coding error for different regions

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If confidence value is configured for predicted values, then reconstructed frame accuracy is improved, but bitstream complexity increases

Engineering Contradiction:
Improvereconstructed frame accuracyVSAvoidbitstream structure
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the confidence information into a separate confidence map that is independently encoded and transmitted in the bitstream. This segmentation allows the confidence data to be processed separately from the image data, enabling improved reconstructed frame accuracy while managing bitstream complexity through structured organization of the encoded data

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4694132A1Coding method, decoding method, and apparatus
Publication Date: 2026.02.11 HUAWEI TECH CO LTD
  • EP4694132A1 patent drawingFigure 1(a)~2(d)
  • EP4694132A1 patent drawingFigure 3
  • EP4694132A1 patent drawingFigure 4

AI summary

This application discloses encoding and decoding methods and apparatuses, applied to the field of image processing technologies. The method includes: obtaining a bitstream including encoded data of a plurality of image frames; decoding the bitstream to obtain motion information of a current frame; obtaining a predicted value of the current frame based on the motion information; decoding the bitstream to obtain first feature information; obtaining a residual of the current frame and a confidence of the predicted value based on the first feature information; and obtaining a reconstructed frame of the current frame based on the predicted value, the residual of the current frame, and the confidence of the predicted value. This can reduce a coding error caused by low accuracy of predicted content in a coding process, effectively enhancing compression performance.