Automated Visual Scanpath Analysis Reveals the Expertise Level of Micro-neurosurgeons

Exploring the effects of expertise on eye movements and visual search behavior of surgeons may help to improve and speed up the training of novices. However, comparing scan patterns to each other is a non-trivial task. This work employs several state-of-the-art, automated scan pattern comparison methods to re-analyze eye-tracking data of neurosurgeons that was captured during the observation of a tumor removal. We evaluate whether these methods can reproduce the findings from the original manual analysis and compare their performance regarding different eye-tracking metrics. Among the considered methods, SubsMatch revealed significant differences between eye movement patterns of expert surgeons and novices for all stimulus images.