Geometry-Aware Visual Odometry for Bronchoscopic Navigation Via High-Gain Observer Fusion
Mohammadreza Kasaei, Francis Xiatian Zhang, Feng Li, Farshid Alambeigi, Kev Dhaliwal, Mohsen Khadem
AI summary
Problem
Conventional visual odometry struggles with texture-poor airway images, specularities, and vanishing-point singularities in tubular anatomy, leading to frequent tracking failures and drift that limit vision-only bronchoscopic navigation in critical care settings.
Approach
The method detects airway lumens to extract vanishing-point cues for stable forward heading estimation, fusing these with looming-based velocity and noisy visual odometry outputs using a bespoke high-gain observer that enforces airway-following kinematic priors.
Key results
- Reduces absolute trajectory error by >50% vs. ORB-SLAM2, LoFTR-VO, DPVO
- Achieves lowest relative pose error across all ex-vivo sequences
- Lumen-derived vanishing-point heading estimation for weak-parallax regions
- High-gain observer fuses orientation, looming velocity, and noisy VO
Why it matters
Enables reliable, CT-free bronchoscopic navigation for critical care and resource-constrained settings where pre-operative imaging or external sensors are unavailable.
Abstract
Navigational bronchoscopy is critical for pul- monary interventions, yet current platforms depend heavily on pre-operative CT or external sensors, limiting their use in critical care and resource-constrained settings. Vision-only navigation offers a scalable alternative, but conventional visual odometry (VO) struggles with texture-poor airway images, specularities, and the vanishing-point singularities of tubular anatomy, leading to frequent tracking failures and drift. We present a geometry-aware VO framework that explicitly lever- ages vanishing-point cues from airway lumens. Detected lumens are back-projected to 3D rays, whose weighted fusion yields a stable forward heading even when parallax cues are absent. This heading, together with looming-based velocity estimates, is fused with noisy VO outputs using a bespoke high-gain observer that enforces airway-following priors and rejects drift. We validate the method on ex-vivo mechanically ventilated human lungs using electromagnetic tracking as ground truth. Compared to state-of-the-art pipelines (ORB-SLAM2, LoFTR- VO, DPVO), our approach reduces absolute trajectory error by more than 50% and achieves the lowest relative pose error across all test sequences.