Synthesizing the image of a 3-D scene as it would be captured by a
camera from an arbitrary viewpoint is a central problem in Computer
Graphics. Given a complete 3-D model, it is possible to render the
scene from any viewpoint. The generation of the model is a tedious,
error-prone and labor intensive task. Here, we propose to bypass the
model generation phase altogether, and to generate images of a 3-D
scene from any novel viewpoint from prestored images. Furthermore,
unlike methods presented so far, we propose to completely avoid
inferring and reasoning in 3-D by using projective invariants derived
from corresponding points in the prestored images. The correspondence
between features is performed off-line in a semi- automated way with a
simple interface. It is then possible to generate wireframe animation
in real time on a standard computing platform. Well understood texture
mapping methods are then applied to realistically render new views
from the prestored ones.
The method proposed here should allow the integration of computer
generated and real imagery and walkthough in realistic virtual
environments. We illustrate our approach on synthetic and real indoor
and outdoor images.
