Motion Adaptive Blur Filter for Video Artifact Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional video compression techniques produce graphical artifacts in videos, especially when rendering high-definition or ultra-high-definition content at high frame rates.

Innovation Solution

A method of filtering graphical frames by downscaling frames to a lower resolution, calculating a motion vector field, and applying a blur filter based on the refined motion vector field to reduce artifacts and simulate motion blur.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If conventional video compression techniques are used, then bandwidth requirements are reduced, but graphical artifacts are produced in the video

Engineering Contradiction:
Improvebandwidth requirementsVSAvoidgraphical artifacts
Core Design Contradiction:
Loss of energyVSObject-generated harmful factors

Solution Approach 1:

The patent applies motion blur filtering as a pre-processing step before video compression. By simulating motion blur on the original high-resolution frames before downsampling and compression, the system prepares the visual content in advance to mask compression artifacts, thereby reducing the harmful effects of compression while maintaining bandwidth efficiency

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent converts the harmful compression artifacts into a benefit by using motion blur as a masking mechanism. The motion blur effect, which naturally softens moving objects, is applied to mask the blocky artifacts produced by compression, thereby transforming the visual impact of compression into an aesthetically acceptable or even desirable motion effect

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

2Manufacturing precision

If high-definition or ultra-high-definition content is rendered at high frame rates, then video quality is improved, but computational resources and bandwidth requirements increase

Engineering Contradiction:
Improvevideo qualityVSAvoidcomputational resources
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent changes the resolution parameter by downsampling high-definition or ultra-high-definition content to lower resolutions after applying motion blur. This parameter transformation allows the system to maintain perceived video quality through motion effects while significantly reducing computational resources and bandwidth requirements for processing and transmission

Inventive Principle:
Principle #35Parameter changes

3Object-generated harmful factors

If motion blur filter is applied to high-resolution frames, then graphical artifacts are reduced, but processing time increases

Engineering Contradiction:
Improvegraphical artifactsVSAvoidprocessing time
Core Design Contradiction:
Object-generated harmful factorsVSLoss of time

Solution Approach 1:

The patent segments the motion blur filtering process by applying it selectively based on motion detection. Instead of uniformly processing all frames at full resolution, the system identifies regions with significant motion and applies motion blur primarily to those segments, thereby reducing the overall processing time while still effectively masking compression artifacts in motion-prone areas

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4042683B1System and method for motion adaptive filtering as pre-process to video encoding
Publication Date: 2025.02.19 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4042683B1 patent drawingFigure 1
  • EP4042683B1 patent drawingFigure 2
  • EP4042683B1 patent drawingFigure 3

AI summary

A method of providing visual information to a user includes receiving a first frame having a first resolution, downscaling the first frame to a second resolution, the second resolution being lower resolution than the first resolution, receiving a second frame having the first resolution, downscaling the second frame to the second resolution, comparing the first frame at the second resolution to the second frame at the second resolution to create a second resolution motion vector field, seeding a seeded first resolution motion vector field, comparing the first frame at the first resolution to the second frame at the first resolution to refine the seeded first resolution motion vector field to create a refined first resolution motion vector field, and applying a blur filter to the first frame at the first resolution based on the refined first resolution motion vector field to create a filtered first frame.