The visual hull of solids of revolution

The visual hull of a 3-D object is a new geometric tool for dealing with concave object recognition or reconstruction using silhouettes. Two equivalent definitions of the visual hull of an object S can be given: (i) the maximal object silhouette-equivalent to S (that gives the same silhouettes as S); (ii) the closest approximation of S that can be obtained with volume intersection techniques. The paper addresses the problem of computing the visual hull of solids of revolution. An algorithm for this purpose is presented and its complexity analyzed. The algorithm operates in an axial plane of S dissected by straight lines and hyperbolae.>