Contents 1. Introduction
Light intensity as a function of space and time defines a hypersurface$$S=(x, y, t, f(x, y, t)) \eqno{\hbox{(1)}}$$ that has the form of a three-dimensional Monge patch. From a geometric point of view the curvature is the most important property of the surface in that it determines the intrinsic structure of the manifold [13]. Geometric methods in computer vision most often deal with the extrinsic geometry of objects in 3D space and how these objects and their motions project on the image plane. However, the geometry of the hypersurface has been used for motion detection [10] with an algorithm based on the gradient of . It has also been shown how the Gaussian curvature of can be used to detect motion discontinuities [18].
