Related Experiment Video
Updated: Jun 28, 2026
06:45
Automated Joint Space Detection Improves Bone Segmentation Accuracy
Published on: November 28, 2025
Bone segmentation and fracture detection in ultrasound using 3D local phase features
Ilker Hacihaliloglu1, Rafeef Abugharbieh, Antony Hodgson
1Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada. ilkerh@ece.ubc.ca
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|November 5, 2008
Summary
This study introduces a new 3D ultrasound method for precise bone segmentation in computer-assisted orthopaedic surgery. The technique accurately identifies bone surfaces and fractures, improving surgical planning and outcomes.
Area of Science:
- Medical Imaging
- Computer-Assisted Surgery
- Biomedical Engineering
Background:
- 3D ultrasound (US) is a promising imaging tool for computer-assisted orthopaedic surgery (CAOS).
- Accurate bone segmentation in 3D US is challenging due to image noise and artifacts.
- Current methods struggle with precise localization of bone surfaces and fracture fragments.
Purpose of the Study:
- To develop an intensity-invariant method for automatic bone segmentation in 3D ultrasound.
- To extract bone surfaces and fractured fragments with high accuracy for CAOS applications.
- To validate the proposed technique in various experimental settings.
Main Methods:
- Utilized 3D Log-Gabor filter banks to extract intensity-invariant local image phase features.
- Extended 2D phase symmetry features to a 3D context for enhanced feature extraction.
- Applied the developed features for automatic segmentation of bone structures in 3D US data.
Main Results:
- Achieved highly accurate segmentation of bone surfaces and fractured fragments.
- Demonstrated localization accuracy better than 0.62 mm for bone surfaces.
- Reported mean errors below 0.65 mm in estimating fracture displacements.
- Validated the technique through phantom, in vitro, and in vivo experiments.
Conclusions:
- The proposed 3D local image phase feature extraction method offers a robust solution for bone segmentation in 3D ultrasound.
- This technique significantly enhances the precision and reliability of imaging for computer-assisted orthopaedic surgery.
- The high accuracy and clinical utility suggest strong potential for improving surgical interventions and patient outcomes.
