The ray-based approach is powerful and useful for the description of 3-D images because of its neutrality to all types of 3-D displaying technologies This approach has been developed in order to reproduce the whole visual data captured at the transmitter However, for the purpose of augmented reality, not only is the reproduction of real visual data essential, but also the synthesis of virtual visual data For example, you may want to change the lightening condition of a scene which has already been captured in some photographs The goal of this work is to synthesize a virtual scene, in which we can easily move light sources The key feature is that the virtual scene is synthesized only from the photographs. From this point of view, this paper extends the concept of the ray-based description of 3-D visual data in order to cope with virtual light sources Experimental results show the potential applicability of this method to the augmented spatial communication systems.
