Gesture-controlled image management for operating room: A randomized crossover study to compare interaction using gestures, mouse, and third person relaying

Abstract

Objective In this work, we aim at comparing formally three different interaction modes for image manipulation that are usable in a surgery setting: 1) A gesture-controlled approach using Kinect 1; 2) oral instructions to a third part dedicated to manipulate the images; and 3) direct manipulation using a mouse. Materials and Methods Each participant used the radiology image viewer Weasis with the three interaction modes. In a crossover randomized controlled trial participants were attributed block wise to six experimental groups. For each group, the order for testing the three modes was randomly assigned. Nine standardized scenarios were used. Results 30 physicians and senior medical students participated in the experiment. Efficiency, measured as time used to pass the scenario, was best when using the mouse (M = 109.10s, SD = 25.96), followed by gesture-controlled (M = 214.97s, SD = 46.29) and oral instructions (M = 246.33s, SD = 76.50). Satisfaction, measured by a questionnaire, was rated highest in the condition mouse (M = 6.63, SD = 0.56), followed by gesture-controlled (M = 5.77, SD = 0.93) and oral instructions (M = 4.40, SD = 1.71). Differences in efficiency and satisfaction rating were significant. No significant difference in effectiveness, measured with error rates, was found. Discussion The study shows with formal evaluation that the use of gestures is advantageous over instructions to a third person. In particular, the use of gestures is more efficient than verbalizing instructions. The given gestures could be learned easily and reliability of the tested gesture-control system is good. Conclusion Under the premise that mouse cannot be used directly during surgery, gesture-controlled approaches demonstrate to be superior to oral instructions for image manipulation.