Related Experiment Videos
Prostate boundary segmentation from 2D ultrasound images
1Imaging Research Laboratories, The John P. Robarts Research Institute, London, Ontario, Canada.
Medical Physics
|September 13, 2000
Summary
This study introduces a semiautomatic prostate segmentation algorithm for cancer treatment planning. The novel method significantly reduces the time and effort required for prostate outlining in ultrasound images, improving efficiency and accuracy.
Area of Science:
- Medical Imaging
- Computational Anatomy
- Oncology
Background:
- Manual prostate segmentation for cancer therapy is time-consuming.
- Accurate prostate outlining is crucial for effective treatment planning.
- Automated segmentation methods can improve efficiency and consistency.
Purpose of the Study:
- To develop and evaluate a semiautomatic algorithm for prostate segmentation from 2D ultrasound images.
- To compare the performance of the semiautomatic method against manual outlining.
- To assess the accuracy and efficiency of the proposed segmentation technique.
Main Methods:
- A model-based initialization approach using four user-selected points.
- Employing cubic interpolation and shape information for initial contour estimation.
- Utilizing an efficient discrete dynamic contour for automatic contour deformation.
- Incorporating contour editing tools for challenging cases.
Main Results:
- The semiautomatic algorithm achieved an average boundary distance of less than 5 pixels (0.63 mm) compared to manual outlining.
- Accuracy and sensitivity for area measurements exceeded 90%.
- The algorithm demonstrated robustness across a range of prostate images.
Conclusions:
- The developed semiautomatic prostate segmentation algorithm offers a significant improvement over manual methods in terms of time and effort.
- The algorithm provides high accuracy and sensitivity, suitable for clinical application in cancer treatment planning.
- This approach enhances the efficiency and consistency of prostate segmentation in 2D ultrasound imaging.