| 摘要 |
<p>Video signals which represent a scene as viewed by a camera are processed to combine a computer generated object with the video signals with the effect that the computer generated object appears within the scene when the video signals are displayed. The scene includes a first object. The process includes mapping a virtual model of the first object to a position of the first object within the scene so that the virtual model substantially corresponds with the real object. The virtual model has a degree of transparency such that the virtual model can be rendered as a substantially transparent virtual object within the scene. The process further includes detecting occluded regions of the virtual model. The occluded regions correspond to regions of the virtual model which are hidden from a virtual light source by the computer generated object. The process also includes modifying the degree of transparency of the virtual model at the occluded regions so as to generate modified transparency regions, and rendering the scene in combination with the computer generated object and the virtual model such that the modified transparency regions of the virtual model appear as shadows on the first object.</p> |
