Related Experiment Video
Updated: Aug 11, 2026
08:52
3D Ultrasound Imaging: Fast and Cost-effective Morphometry of Musculoskeletal Tissue
Published on: November 27, 2017
Prostate boundary segmentation from 3D ultrasound images
Ning Hu1, Dónal B Downey, Aaron Fenster
1Imaging Research Laboratories, The John P. Robarts Research Institute, London, Ontario N6H 5C1, Canada.
Medical Physics
|August 9, 2003
Summary
This study presents a fast, semiautomatic algorithm for prostate segmentation in 3D ultrasound images, requiring minimal user input for accurate boundary estimation and refinement.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Urology
Background:
- Accurate prostate boundary segmentation is crucial for prostate cancer management.
- Existing segmentation methods may be time-consuming or require extensive user interaction.
Purpose of the Study:
- To develop and evaluate a semiautomatic algorithm for prostate segmentation from 3D ultrasound images.
- To improve the efficiency and accuracy of prostate boundary delineation.
Main Methods:
- A model-based initialization using six user-selected points.
- Mesh refinement with an efficient deformable model.
- An interactive editing tool for boundary correction and re-deformation.
Main Results:
- The algorithm achieves segmentation in under 1 minute.
- Global error analysis showed an average absolute difference of 1.19 +/- 0.14 mm and an average volume difference of 7.16% +/- 3.45% compared to manual segmentation.
- Algorithm segmentation variability (0.63 mm) was lower than manual segmentation variability (0.98 mm).
Conclusions:
- The developed semiautomatic algorithm provides efficient and accurate prostate segmentation from 3D ultrasound.
- The method offers a promising tool for clinical applications in prostate cancer management.
- The algorithm demonstrates reduced variability compared to manual segmentation, enhancing reliability.
