Dynamic real-world scaling for enhanced PS-FLIm data visualization in TORS
Pu Sun, Mohamed Hassan, Alexander E. Taffe, Suparn Sathya, Katjana Ehrlich, Julien Bec, Laura Marcu · 2025
This project enhances real-time guidance in Transoral Robotic Surgery (TORS) using Pulse-Sampling Fluorescence Lifetime Imaging (PS-FLIm). We describe a dynamic method that adjusts the interpolation disk size based on the aiming beam probe's dimensions and instance segmentation mask, ensuring accurate tissue mapping. By calculating the scale in dots per millimeter (dpmm) and incorporating a Gaussian-weighted function, our approach emphasizes closer data points and reduces the influence of distant ones. Preliminary results show significant improvements compared to the fixed inverse distance method, validated with surgical and ex-vivo videos, achieving an error rate of just 1.2 mm, thus enhancing the reliability of FLIm data visualization.