Volumetric Segmentation in Planar Medical Images with Radiologist Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging tools are slow and inefficient in providing quantitative measurements and volumetric segmentation of structures, relying heavily on manual delineation by radiologists, which is time-consuming and not suitable for fast-paced clinical workflows.
Innovation Solution
A semi-automated segmentation process that allows radiologists to interactively define and visualize long and short axes, using a system that determines volumetric segmentation by combining user input with statistical sampling and probability distributions to classify voxels into foreground and background classes, enabling rapid and accurate 3D contour generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual delineation by radiologists is used, then measurement precision is improved, but productivity deteriorates
Solution Approach 1:
The system enables semi-automated segmentation where the algorithm performs self-service by automatically classifying voxels into foreground and background classes using statistical sampling and probability distributions, reducing reliance on manual radiologist delineation while maintaining accuracy
Solution Approach 2:
The system changes the operational parameters by using statistical sampling and probability distributions to classify voxels, transforming the manual delineation process into an automated computational process that maintains precision while improving productivity
2Productivity
If full automated segmentation is implemented, then productivity is improved, but measurement precision deteriorates
Solution Approach 1:
The system incorporates feedback mechanisms where the semi-automated process allows radiologists to review and correct algorithm-generated segmentations, ensuring measurement precision is maintained while benefiting from the productivity gains of automation
Solution Approach 2:
The system applies partial automation rather than full automation, using statistical sampling to classify voxels while retaining radiologist oversight for critical decisions, achieving optimal balance between productivity and precision
3Measurement precision
If manual slice-by-slice contouring is used, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary automated classification of voxels into foreground and background classes using statistical sampling, preparing the segmentation framework in advance before radiologist review, thereby reducing the time radiologists spend on manual contouring while maintaining boundary accuracy
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems for volumetric segmentation of structures in planar medical images. One example method includes displaying a first planar medical image. The method farther includes receiving a user input indicating a line segment in the first planar medical image. The method also includes determining an inclusion region using the line segment. The inclusion region consists of a portion of the structure. The method further includes determining a containment region using the line segment. The containment region includes the structure. The method also includes determining a background region using the line segment. The background region excludes the structure. The method further includes determining a three dimensional (3D) contour of the structure using the inclusion region, the containment region, and the background region. The method also includes determining a long axis of the structure using the 3D contour. The method further includes outputling a dimension of the long axis.