3D Virtual Reality Navigation by Human's Head Movements Using a Generic Webcam

摘要

This paper presents an interactive interface in order to provide the user with 3D experience of virtual reality navigation. The applied computer vision sensor is a generic mono webcam whose resolution supports at least 320x240 pixels. Practically the head movements are acquired, analyzed and translated to adapt the virtual reality to the relative angle of view in a real time progress. The head tracking method is based on face detection using two skin colour models, one in YUV (YCbCr) and the other in RGB colour space. Face containing images acquired by webcam are then classified using intersection of the two skin models' output. Afterwards the tracked head position and scale values are translated into rotations and zooming factors respectively to update the virtual reality viewport.