Look2React: Making VR NPCs Come Alive with Dynamic Vision-Guided Reactions (Conditionally Accepted to IEEE VR 2026)
Published in The 33rd IEEE Conference on Virtual Reality and 3D User Interfaces (VR ’26), 2025
We present Look2React, an interaction system that captures what NPCs see, using a vision-based reasoning model to select pose and text responses. As such, Look2React endows NPCs with the ability to react dynamically and appropriately to player interactions.
Recommended citation: Ritik Vatsal, Xincheng Huang, and Robert Xiao. 2025. In 33rd IEEE Conference on Virtual Reality and 3D User Interfaces (VR ’26). (To Appear) https://xincheng.me/publications/look2react
