Heart Lumen Segmentation Error Detection in Ultrasound Imaging

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing ultrasound diagnostic systems require significant time for users to manually check multiple frames to correct segmentation errors in heart lumen images due to artifacts, leading to inefficiencies in calculating ejection fraction.

Innovation Solution

An ultrasound diagnostic apparatus that includes an image acquisition unit, contour extraction unit, cross-sectional area calculation unit, cross-sectional area change curve generation unit, and error detection unit to automatically detect and correct segmentation errors by comparing actual and predicted cross-sectional area change curves, allowing for quick identification and correction of errors using control points set by the user.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the user manually checks multiple ultrasound frames to find segmentation errors, then the segmentation accuracy can be verified, but the time required for correction increases significantly

Engineering Contradiction:
Improvesegmentation accuracyVSAvoidcorrection time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system generates a cross-sectional area change curve from segmented contours and compares it with a prediction curve to automatically detect segmentation errors. This feedback mechanism allows the system to identify errors without requiring manual inspection of multiple frames, thus maintaining segmentation accuracy while significantly reducing correction time

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The error detection unit automatically identifies segmentation errors by comparing the actual cross-sectional area change curve with the predicted curve, enabling the system to self-diagnose issues without user intervention. This eliminates the need for users to manually check multiple frames, resolving the contradiction between verification accuracy and time consumption

Inventive Principle:
Principle #25Self-service

2Reliability

If the system processes multiple frames to ensure accurate segmentation, then the reliability of ejection fraction calculation improves, but the processing time increases

Engineering Contradiction:
Improveejection fraction calculation reliabilityVSAvoidprocessing speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system uses feedback from cross-sectional area change curves to automatically detect segmentation errors across multiple frames. By comparing actual curves with predicted curves, the system can reliably identify errors without requiring extensive manual processing of each frame, thus maintaining calculation reliability while improving processing speed

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary automatic error detection using cross-sectional area change curve comparison before final ejection fraction calculation. This preliminary action identifies potential errors early, allowing for efficient correction and ensuring reliable results without the need for time-consuming manual review of all frames

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4712021A1Ultrasound diagnostic apparatus and method of controlling ultrasound diagnostic apparatus
Publication Date: 2026.03.18 FUJIFILM CORP
  • EP4712021A1 patent drawingFigure 1
  • EP4712021A1 patent drawingFigure 2~3
  • EP4712021A1 patent drawingFigure 4~5

AI summary

Provided are an ultrasound diagnostic apparatus and a control method of the ultrasound diagnostic apparatus capable of easily correcting a segmentation result of a lumen of a heart in a short time. An ultrasound diagnostic apparatus includes: an image acquisition unit that acquires ultrasound images of a plurality of frames that are continuous in time series and in which a heart of a subject is imaged; a contour extraction unit that performs segmentation on the ultrasound images of the plurality of frames to extract a contour of a lumen of the heart for each frame; a cross-sectional area calculation unit that calculates a cross-sectional area of the lumen of the heart for each frame based on the extracted contour; a cross-sectional area change curve generation unit that generates a cross-sectional area change curve representing an actual temporal change of the cross-sectional area based on calculated the cross-sectional area; and an error detection unit that detects an error of the performed segmentation by comparing the generated cross-sectional area change curve with a predicted cross-sectional area change prediction curve.