Related Experiment Videos
Automatic needle segmentation in three-dimensional ultrasound images using two orthogonal two-dimensional image
Mingyue Ding1, H Neale Cardinal, Aaron Fenster
1Robarts Research Institute, London, ON N6A 5K8, Canada.
Medical Physics
|February 28, 2003
Summary
This study presents an algorithm for segmenting needles in 3D ultrasound images using two 2D projections. The method achieves near real-time performance with high accuracy for needle localization.
Area of Science:
- Medical Imaging
- Computer Vision
- Biomedical Engineering
Background:
- Accurate needle segmentation in 3D ultrasound is crucial for minimally invasive procedures.
- Existing methods often struggle with real-time performance and complex imaging environments.
Purpose of the Study:
- To develop and evaluate an algorithm for robust 3D needle segmentation from ultrasound images.
- To reduce the complexity of 3D segmentation to two 2D segmentation tasks.
Main Methods:
- Algorithm utilizes two orthogonal 2D projections of a 3D ultrasound volume.
- Employs volume cropping and Gaussian transfer functions for background noise reduction.
- Reduces 3D needle segmentation to two 2D needle segmentations.
Main Results:
- Algorithm achieves near real-time performance (approx. 10 fps).
- Demonstrates high accuracy with root-mean-square error < 0.8 mm for needle length and endpoint coordinates.
- Average accuracy of ~0.5 mm achieved for needle lengths ranging from 4.0 mm to 36.7 mm.
Conclusions:
- The proposed algorithm offers an efficient and accurate solution for 3D needle segmentation in ultrasound imaging.
- The method's reliance on 2D projections simplifies the segmentation process.
- Suitable for applications requiring real-time needle tracking and guidance.