3D shape recovery across two views based on approximate geometric constraints

We propose a new approach to object matching and shape reconstruction across two views. The approach relies on the topological relationship among the feature points and a crude rigidity constraint (epipolar condition) in carrying the object from the model view to the scene view to a first good approximation. The correspondences are then refined by fitting an active contour model to the transferred feature points on the scene view, which automatically assumes the shape of the object on correct matching. Under the assumption of weak perspective projection, it can be shown that the establishment of such correct correspondences results in the full determination of the relative depths of all the matched object points in 3D space and thus recovery of the complete 3D structure of the object.