Eye-gaze Based Attention Estimation in Minimal Invasive Surgery
Sprache des Vortragstitels:
Englisch
Original Tagungtitel:
ECEM 2015 (18th European Conference on Eye Movements)
Sprache des Tagungstitel:
Englisch
Original Kurzfassung:
We hypothesize that laparoscopic surgical suboptimal outcomes in patient safety measures are correlated with (i) cognitive load / level of attention of the operating surgeon, (ii) the frequency and degree of disruptions to the surgical workflow, and (iii) the misalignment of visual and motor axes in laparoscopic equipment / setting (eye-hand coordination). In this paper we present an eye tracking approach to assess the surgeons attention based on a clinical study with 6 laparoscopic surgeries. At the eye-gaze level an efficient way for filtering saccades over tremors is presented, and a strategy for reducing the complexity of data by methodical discretizing fixations over saccades while a preserving maximum of the semantic information is presented. This procedure was automated and videos were generated to estimate the surgeons attention levels and to identify possible causes of distraction.