Human-centric indoor environment modeling from depth videos

Abstract

We propose an approach to model indoor environments from depth videos (the camera is stationary when recording the videos), which includes extracting the 3-D spatial layout of the rooms and modeling objects as 3-D cuboids. Different from previous work which purely relies on image appearance, we argue that indoor environment modeling should be human-centric: not only because humans are an important part of the indoor environments, but also because the interaction between humans and environments can convey much useful information about the environments. In this paper, we develop an approach to extract physical constraints from human poses and motion to better recover the spatial layout and model objects inside. We observe that the cues provided by human-environment intersection are very powerful: we don't have a lot of training data but our method can still achieve promising performance. Our approach is built on depth videos, which makes it more user friendly. © 2012 Springer-Verlag.