Computer Vision enhanced Advanced Cone Beam CT for Intraoperative Guidance

Project Overview

Cone Beam Computed Tomography (CBCT) is a valuable imaging technique used in various medical procedures, particularly in radiation therapy and interventional procedures. Our research focuses on enhancing CBCT with computer vision and machine learning techniques to improve intraoperative guidance during bronchoscopy procedures.

The Challenge

Current image-guided bronchoscopy systems face challenges in accurately tracking the bronchoscope's position within the complex bronchial tree, especially when dealing with respiratory motion and tissue deformation. This can lead to navigation errors and reduced procedural efficiency.

Our Approach

We are developing an advanced system that combines:

• 4D-CT imaging to capture respiratory motion patterns
• Machine learning algorithms to model and predict tissue deformation
• Real-time computer vision techniques to enhance bronchoscopic navigation
• Registration methods that align pre-procedure CT images with intraoperative bronchoscopic video

Expected Outcomes

Our enhanced CBCT system aims to:

• Improve the accuracy of bronchoscope tracking within the airways
• Reduce procedure time and increase diagnostic yield
• Enhance the physician's ability to reach peripheral lung lesions
• Provide more precise guidance for biopsy and treatment delivery

Current Status

This project is currently in the development and validation phase. We are working with clinical partners to test our system in simulated environments before moving to clinical trials.