Augmenting Vision with Co-Located Live Thermal View in Mixed Reality

Integrating live sensory extensions, such as thermal imagery, into Mixed Reality (MR) presents a significant opportunity to augment human vision. However, the perceptual and ergonomic costs of interacting with such data remain underexplored. This study introduces a mixed-reality system that integrates a live thermal feed from a Teledyne FLIR Hadron 640R sensor with a high-precision Varjo XR-3 headset that allows eye and hand tracking. We compared two interaction modalities to access the augmented sensory information: a gaze-based "Gaze-Enlarge"interface and a gesture-based "Hand-Menu Zoom"in an experiment with 24 participants involving three thermal inspection tasks with varying cognitive loads. Mixed-effects analysis revealed benefits for the more direct gaze-based method over the more command-like gesture-based method: while both modalities demonstrated comparable accuracy and task efficiency, gaze interaction incurred substantially lower perceived workload and physical demand. Participants consistently rated gaze control as significantly faster, easier, and more natural. Furthermore, we identified a key adaptive strategy: under high cognitive load, participants strategically shifted from continuous hand-slider adjustments (used for search) to discrete button presses, prioritizing interaction stability. These findings identify a complementary division of labor with gaze used for rapid selection and hand for explicit confirmation, and establish design principles for efficient augmented perception interfaces. All software, hardware configurations, and 3D models are openly released to support reproducible research.