Use of Image Segmentation Algorithm to Test Model Spectral Distribution for Surgical Lighting

Lighting is an essential tool in surgery. Altering the spectral distribution of the lighting can heighten visual contrast between anatomical features of interest and the surrounding tissues. Due to the difficulty in assessing lighting spectral distributions on surgical scenes, we use image segmentation algorithm as a means of testing the relative merits of model spectral distribution for surgical lighting, compared to conventional surgical lighting. The relative accuracy in identifying the common bile duct (CBD), as known from the annotated collected hyperspectral images of surgical scenes, is used as a figure of merit to determine viability of lighting spectral