A Metrological Approach to the Performance Characterization of Eye- and Head-Tracking Interfaces in eXtended Reality

Given the increasing adoption of eXtended Reality Head-Mounted Displays (XR HMDs) and the related demand for hands-free Human-Machine Interaction (HMI), this paper introduces a metrological approach to the performance charac-terization of Eye- and Head-tracking input interfaces within XR scenarios. In particular, two metrics of interest are identified: Accuracy in selecting objects and Euclidean Distance between the detected gaze and reference points. Without loss of generality, the Microsoft HoloLens 2 HMD was considered as a case study. An experimental campaign involving eight subjects was conducted to investigate the intra- and inter-individual variability of the metrics considered. The obtained experimental results showed robust performance within the interfaces of HoloLens 2, thus allowing the development of high-information-rate XR applications in scenarios where precision tasks are needed.