Abstract
This paper proposes a novel technique for recovering entire shape of a solid simultaneously by the employment of virtual see-through cameras. In the technique, a solid is surrounded by a set of cameras. The rear shape of the solid is transformed into the shape which would be observed from frontal cameras and it is superposed to the frontal shape, and vice versa. All this procedure results in that the entire shape of the solid is observed by all the cameras surrounding it. This gives stronger geometrical constraints in the shape recovery. The factorization is applied to this information and the entire shape recovers simultaneously. Experimental results show that this technique gives higher precision in the recovered shape than the technique proposed before by the authors.
