A 3D simulation approach for multi-human detection using a multi-sensor frame

Abstract

In this work, we introduce a 3D simulation framework for multi-human detection using multi-sensor. We build a 3D indoor space by modeling the real dimensions of a classroom using the 3D creation tool Blender. We also create human animations like sitting down and sleeping using its game engine. The multi-sensor frame consists of two RGB and two depth cameras simulated with the aid of the simulator MORSE. In addition, we examine the feasibility of simulation by testing the baseline vision method of face and human detection on a textured human animation. ？ 2013 IEEE.