Human Proxemics Behaviour

Human proxemics behavior, the study of spatial relationships in human interaction, is a growing area of research focusing on understanding and modeling how people position themselves relative to others and robots. Current work employs deep learning models, including variational autoencoders and diffusion models, to reconstruct and predict proxemics from visual data, often incorporating emotional context and physical constraints to improve accuracy. This research is significant for advancing human-robot interaction, particularly in designing socially assistive robots that respect personal space, and for improving the accuracy of 3D human pose estimation in complex social scenarios.