LARK provides low cost accurate tracking for image guided surgery
LARK: A Low-Cost, Accurate, Occlusion-Resilient, Kalman Filter-Assisted Tracking System for Image-Guided Surgery
Computer Vision and Pattern Recognition
Summary
Surgical tools need to be tracked precisely during operations so doctors know where they are inside the body. Existing expensive tracker systems can lose track when something blocks their view, making procedures harder. The authors made LARK, a system using cheaper regular cameras and smart software to keep tracking tools accurately even if some cameras are blocked. Their system nearly matches the accuracy of costly trackers and works well with multiple cameras sharing data. This helps make image-guided surgery more accessible and affordable.
What this means in practice
- •For surgical equipment developers: Build affordable multi-camera tracking systems that maintain accuracy despite occlusions during surgeries using commodity RGB cameras and Kalman filtering.$Commercial implications: Enables development of low-cost optical tracking products for surgical navigation that compete with expensive stereo infrared systems by maintaining accuracy and occlusion resilience.
- •For robotic surgery system builders: Integrate LARK’s multi-view fusion and adaptive filtering methods to improve instrument tracking robustness in robot-assisted surgical procedures.
Authors
George Sideris, Justin Cree, Andrew Stirling, Mamadou Ly, Étienne Léger, D. Louis Collins
Abstract
Image-guided surgery (IGS) depends on accurate tracking of surgical instruments to provide real-time navigation relative to anatomical structures. Commercial stereo infrared trackers are accurate but prone to occlusion and cost-prohibitive for many settings. This work presents LARK, a multi-camera optical tracking system using commodity RGB hardware and multi-view redundancy and fusion. We develop and evaluate two complete tracking methods: multi-view monocular pose fusion and multi-view triangulation. Both methods are assessed under varying occlusion levels using a precision-machined grid and an anatomical head phantom, and compared against a gold-standard stereo infrared system. With five cameras and adaptive Kalman filtering, LARK achieves median target registration errors of 0.64 mm for point localization with triangulation and 0.73 mm for trajectory tracking with pose fusion on the machined grid. Camera-subset experiments show graceful degradation in adaptive pose-fusion accuracy as fewer views remain available. With tracking hardware costing under $1,000 USD, LARK provides a low-cost platform for image-guided surgery research. Hardware designs and software are publicly available at https://nist.mni.mcgill.ca/software/ , and datasets at https://nist.mni.mcgill.ca/data/ .