Skill Measurement via Egocentric Vision in Wetlab

With the development in egocentric vision, skill measurement has been recently proposed as a novel topic in this emerging field. In this report, we record the the experimenter’s first-person videos in Wet labratories (wetlab), and measure his/her operative skills. Specifically, given the videos of expert and amateur, we analyze their head motions, hands motions, eye-hand coordinations and key-moment actions. Accordingly, we investigate the cues that are discriminative between amateur and expert. The analytical results show that during the experiment expert is more likely to involve faster head/hands movements, more consistent eye-hand manipulation and more stable pipetting actions compared with amateur.