Ultrasound Motion Estimation for Artifact Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional ultrasound imaging techniques face challenges in increasing image resolution, eliminating artifacts, and suppressing speckle, especially in compound imaging due to motion between frames, which is exacerbated by poor signal-to-noise ratio and large flat regions, making motion estimation difficult.

Innovation Solution

A method for estimating motion between images by calculating global and local motion vectors using sum of absolute differences or normalized cross-correlation, adjusting functions with terms dependent on distance and unnatural motion detection, to generate motion-compensated compound images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If spatial compounding is used to reduce speckle and improve image quality, then image quality improves, but motion artifacts increase due to frame-to-frame movement

Engineering Contradiction:
Improveimage qualityVSAvoidmotion artifacts
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent applies motion estimation and compensation techniques before the compounding process to predict and correct for motion between frames. By estimating motion vectors from reference frames and applying compensations in advance, the system prevents motion artifacts from forming in the final compounded image, thereby maintaining high image quality without the harmful effects of motion distortion.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If conventional block matching motion estimation is used, then motion compensation can be applied, but accuracy is insufficient due to poor signal-to-noise ratio and large flat regions in ultrasound images

Engineering Contradiction:
Improvemotion estimation accuracyVSAvoidpoor signal-to-noise ratio
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent divides the image into multiple blocks of different sizes and performs motion estimation at multiple resolution levels. This hierarchical segmentation allows the algorithm to first estimate coarse motion at low resolution, then refine the estimates at higher resolutions. This approach improves accuracy by breaking down the complex motion estimation problem into manageable segments that are less sensitive to noise and flat regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a temporal dimension by using motion compensation across multiple frames in addition to spatial information. By incorporating temporal coherence and using motion vectors from previous frames as priors, the system enhances motion estimation accuracy in the presence of poor signal-to-noise ratio and large flat regions that challenge conventional spatial-only methods.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Reliability

If multiple component frames are acquired from different spatial directions, then speckle reduction is achieved, but acquisition time increases due to limited frame rate

Engineering Contradiction:
Improvespeckle reductionVSAvoidacquisition time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies motion compensation techniques to align frames before compounding, which allows the system to use fewer component frames while achieving the same speckle reduction effect. By pre-aligning frames using estimated motion vectors, the system reduces the need to acquire many frames to achieve proper alignment, thereby reducing acquisition time while maintaining speckle reduction benefits.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3686843A1Method and system for estimating motion between images, particularly in ultrasound spatial compounding
Publication Date: 2020.07.29 ESAOTE
  • EP3686843A1 patent drawingFigure 1
  • EP3686843A1 patent drawingFigure 2
  • EP3686843A1 patent drawingFigure 3

AI summary

Methods are provided for estimating motion between images associated with a common region of interest, the method comprising: providing frames including a reference frame and a target frame; determining a global motion vector based on a comparison of the reference and target frames; for a plurality of local blocks, determining local motion vectors between the reference and target frames based on the global motion vector to form globally adjusted local motion vectors; considering the globally adjusted local motion vectors as motion estimator. A corresponding system is also disclosed.