Computational biomodel of motion parallax for multiview 3D video conferencing

The brain makes use of multiple visual cues to reconstruct the three-dimensional structure of a scene. Neuroscience studies have demonstrated that motion parallax is one such powerful depth cue. Motion parallax consists in harnessing the motion of the observer, offering a different view of the environment depending on his/her position to get some 3D feeling. This paper introduces a biologically inspired computational model of motion parallax, seeking for a novel application to multiview 3D video conferencing. Indeed, video conference systems have evolved little regarding 3D vision for viewing video apart from using new 3D technology, making use of special glasses and auto-stereoscopic screens. Our challenge is to take advantage of neuronal visual cues that can be used to provide 3D perception apart from stereoscopy. Various composition models (change modes between cameras) have been designed and tested in order to analyze which one provides a better 3D effect.