Abstract
The reconstruction of a 3-D moving object from its images is discussed. Inequalities describing the relation between the required number of points observed on the object and the number of frames are given, and the scale and orientational invariance of the procedure are discussed.