Global Registration Algorithm for Bronchoscopy Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image-based bronchoscopy guidance systems face challenges such as manual registration errors, need for technician assistance, and inability to detect and correct faulty bronchoscope maneuvers, especially during patient movements like coughing, which affect navigation accuracy and efficiency.
Innovation Solution
A global-registration algorithm and a semi-global solution that provide updated navigational information and guidance by precomputing bronchoscopy-friendly orientations along a planned route, allowing for automatic detection and correction of faulty maneuvers and adverse events, enabling single-user control without a technician.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual registration is used to align bronchoscope position with 3D MDCT space, then the system can provide navigation guidance, but navigation errors occur as early as the second airway generation due to skill variations and mental registration difficulties
Solution Approach 1:
The system automatically performs registration by comparing real bronchoscopic video frames with virtual bronchoscopic images generated from 3D MDCT data. The algorithm autonomously identifies bifurcations and matches them between real and virtual views without requiring manual technician intervention, thereby eliminating skill variation and improving navigation accuracy while reducing operational complexity
Solution Approach 2:
The patent replaces manual mechanical registration operations with an automated image-based algorithm. Instead of technicians mentally aligning coordinates, the system uses computer vision to automatically register the bronchoscope position by matching visual features between real and virtual bronchoscopic views, substituting human cognitive effort with automated image processing
2Reliability
If discrete local registrations at consecutive bifurcations are used to guide the physician, then navigation can be provided, but the system cannot detect and correct faulty bronchoscope maneuvers across multiple bifurcations
Solution Approach 1:
The system precomputes virtual bronchoscopic images for all bifurcations along the planned route to the ROI before the procedure begins. This preliminary preparation creates a complete reference map that enables the algorithm to detect and correct faulty maneuvers by comparing real-time video frames against the precomputed virtual views at each bifurcation, improving reliability without adding procedural complexity
Solution Approach 2:
The system continuously compares real bronchoscopic video frames with corresponding virtual bronchoscopic images and provides feedback on bronchoscope position and orientation. When faulty maneuvers are detected through mismatched features, the system can identify the deviation and guide correction by indicating the correct bifurcation location, thereby improving detection capability while maintaining manageable system complexity through iterative visual feedback
3Measurement precision
If EM-based bronchoscopy guidance is used to track bronchoscope position, then global position can be known in MDCT space, but localization errors occur due to patient breathing motion and the system is susceptible to local EM-field distortions
Solution Approach 1:
The patent replaces EM-based tracking with an image-based registration system that uses visual feature matching between real and virtual bronchoscopic views. This substitution eliminates susceptibility to EM-field distortions and breathing motion interference, as the system relies on optical image comparison rather than electromagnetic field tracking, thereby improving position accuracy while avoiding the harmful effects of EM-based methods
4Extent of automation
If an attending technician must carefully follow the bronchoscope position to provide guidance, then navigation can be maintained, but the system requires considerable special hardware and cannot operate without technician assistance
Solution Approach 1:
The system performs automatic registration and navigation guidance without requiring an attending technician to follow the bronchoscope position. The algorithm autonomously processes real video frames, compares them with precomputed virtual images, identifies bifurcations, and determines bronchoscope location, thereby achieving high automation level while using standard bronchoscopy equipment without considerable special hardware
Data Source
AI summary
Two system-level bronchoscopy guidance solutions are presented. The first incorporates a global-registration algorithm to provide the physician with updated navigational and guidance information during bronchoscopy. The system can handle general navigation to a region of interest (ROI), as well as adverse events, and it requires minimal commands so that it can be directly controlled by the physician. The second solution visualizes the global picture of all the bifurcations and their relative orientations in advance and suggests the maneuvers needed by the bronchoscope to approach the ROI. Guided bronchoscopy results using human airway-tree phantoms demonstrate the potential of the two solutions.