Related Experiment Video
Updated: Jun 8, 2025
04:47
Author Spotlight: Learning Systematic Bronchoscopy in a Simulation-Base Setting
Published on: June 23, 2023
2.3K
BronchoTrack: Airway Lumen Tracking for Branch-Level Bronchoscopic Localization.
IEEE Transactions on Medical Imaging
|November 7, 2024
Summary
BronchoTrack achieves accurate, real-time bronchoscope localization using efficient lumen detection and multi-object tracking. This framework generalizes across patients, enabling precise navigation within complex airway structures for improved interventions.
Area of Science:
- Medical Imaging
- Robotics
- Computer Vision
Background:
- Accurate real-time bronchoscope localization is critical for interventional pulmonology.
- Existing vision-based methods face challenges in balancing speed and generalization.
Purpose of the Study:
- To develop BronchoTrack, a real-time framework for accurate branch-level bronchoscope localization.
- To improve lumen detection, tracking, and airway association for enhanced navigation.
Main Methods:
- Utilized a lightweight detector for efficient lumen detection.
- Introduced multi-object tracking to mitigate temporal confusion during localization.
- Developed a training-free association method using a semantic airway graph for generalization.
Main Results:
- Achieved 81.72% localization accuracy on 11 patient datasets, reaching the 6th generation airways.
- Successfully localized the bronchoscope to the 8th generation airway in an in-vivo porcine model.
- Demonstrated real-time performance with high accuracy and generalization capabilities.
Conclusions:
- BronchoTrack offers a robust solution for real-time bronchoscope localization.
- The framework shows significant potential for clinical applications in interventional pulmonology.
- The combination of efficient detection, tracking, and graph-based association ensures reliable navigation.
