Adaptive Sharpening in Motion Compensation for Artifact Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies suffer from artifacts caused by motion interpolation and quantization, leading to reduced quality in interpolated frames, particularly edge blurring and ringing artifacts, which are not effectively addressed by current sharpening filters.

Innovation Solution

Implement an adaptive sharpening filter after the motion interpolation step, controlled by a sharpening strength coefficient, to reduce artifacts and enhance edge quality, using non-linear filters to adapt to local video content features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If motion interpolation filtering is applied to generate fractional-pel prediction blocks, then motion compensation precision is improved, but ringing artifacts and edge blurring are introduced

Engineering Contradiction:
Improvemotion compensation precisionVSAvoidringing artifacts and edge blurring
Core Design Contradiction:
Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

An adaptive sharpening filter is introduced as an intermediary processing stage between motion interpolation and prediction block generation. This filter acts as a mediator that counteracts the blurring and ringing effects of the interpolation filter while preserving the enhanced motion compensation precision, thereby resolving the contradiction between precision improvement and artifact introduction

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent applies parameter changes by dynamically adjusting the sharpening strength coefficient based on local video content characteristics. By changing the filtering parameters adaptively rather than using fixed parameters, the system can enhance edges and reduce artifacts without compromising the motion compensation precision achieved through fractional-pel interpolation

Inventive Principle:
Principle #35Parameter changes

2Shape

If conventional sharpening filters are applied to reduce artifacts, then edge sharpness is improved, but the filters fail to adapt to local video content features

Engineering Contradiction:
Improveedge sharpnessVSAvoidadaptability to local video content
Core Design Contradiction:
ShapeVSAdaptability or versatility

Solution Approach 1:

The patent transforms the static conventional sharpening filter into a dynamic adaptive filter whose characteristics change based on local video content. The filter adapts its sharpening strength and behavior according to local features such as edge orientations, activity levels, and content types, enabling it to effectively enhance edge sharpness while maintaining adaptability to diverse video content

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The adaptive sharpening filter implements local quality by applying different filtering characteristics to different regions of the video content. By analyzing local features and adjusting filter parameters accordingly, the system applies appropriate sharpening strength to each local area, preserving fine details in some regions while avoiding over-sharpening in others, thus achieving both edge sharpness improvement and content adaptability

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3348057B1Apparatus and method for video motion compensation
Publication Date: 2025.11.26 HUAWEI TECH CO LTD
  • EP3348057B1 patent drawingFigure 1
  • EP3348057B1 patent drawingFigure 2
  • EP3348057B1 patent drawingFigure 3

AI summary

The present invention relates to a video coder for predictive coding a video stream of subsequent frames according to motion compensation into an encoded video bit stream, comprising a frame buffer (207) adapted to store at least one reference frame of the video stream, said reference frame being different from a current frame of the video stream, an inter prediction unit (210) adapted to generate a prediction block of a current block of the current frame from a reference block of the reference frame, and an adaptive sharpening filter (211) configured to adaptively filter the prediction block.