3D reconstruction and virtual forming in rotationally symmetric space

For reconstructing 3D objects and for generating arbitrary views of the scene, accurate calibration of multiple cameras is very important. We introduce rotationally symmetric space, and show that in rotationally symmetric space, cameras can be calibrated and 3D points can be reconstructed just from a single basis line reliably. The proposed method is applied for deforming virtual clay in rotationally symmetric space and for generating augmented reality images of virtual pottery.