Related Experiment Video
Updated: Sep 6, 2025
06:08
A Cognitive Fusion-guided Prostate Biopsy Using Multiparametric Magnetic Resonance Imaging and Transrectal Ultrasound
Published on: March 21, 2025
366
Semi-Automatic Prostate Segmentation From Ultrasound Images Using Machine Learning and Principal Curve Based on
Tao Peng1,2, Caiyin Tang3, Yiyun Wu4
1Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong, Hong Kong SAR, China.
Frontiers in Oncology
|June 24, 2022
Summary
Accurate prostate segmentation in transrectal ultrasound (TRUS) is improved by the Hybrid Segmentation Model (H-SegMod). This novel approach enhances Region of Interest (ROI) segmentation, outperforming existing methods.
Area of Science:
- Medical Imaging
- Biomedical Engineering
- Computer Vision
Background:
- Prostate segmentation in transrectal ultrasound (TRUS) is difficult due to low image contrast and artifacts.
- Accurate segmentation is crucial for diagnosis and treatment planning.
Purpose of the Study:
- To develop a semi-automatic model, Hybrid Segmentation Model (H-SegMod), for improved prostate Region of Interest (ROI) segmentation in TRUS images.
- To address challenges posed by low contrast and imaging artifacts in TRUS.
Main Methods:
- A two-stage cascaded model (H-SegMod) was proposed.
- Stage 1: Improved principal curve-based model using radiologist-selected seed points for vertex sequence generation.
- Stage 2: Improved machine learning model to define a smooth prostate contour map function.
Main Results:
- H-SegMod demonstrated superior segmentation performance compared to state-of-the-art models.
- Achieved high average scores: Dice Similarity Coefficient (DSC) of 96.5%, Jaccard Similarity Coefficient (Ω) of 95.2%, and Accuracy (ACC) of 96.3%.
Conclusions:
- The proposed H-SegMod offers a robust and accurate solution for prostate segmentation in TRUS.
- The model effectively overcomes common imaging challenges, providing reliable segmentation results.
