Summary: Carved Visual Hulls for High­Accuracy Image­Based Modeling
Yasutaka Furukawa (yfurukaw@uiuc.edu) and Jean Ponce (ponce@cs.uiuc.edu)
Beckman Institute, University of Illinois at Urbana­Champaign, Urbana, IL 61801, USA
Figure 1: Constructing the carved visual hull of a toy dinosaur. From left to right: one of the 24 input photographs, the raw visual hull, the
rims found on its surface, the carved visual hull after graph cuts, and the final 3D model after iterative refinement.
The relative accuracy of high­end laser range scanners can be
as high as 1/10,000, allowing the construction of high­accuracy
solid models of complex shapes from registered depth maps [Levoy
et al. 2000]. Comparable accuracy levels can be achieved using
``ordinary'' cameras and sophisticated photogrammetric methods,
but these typically output a relatively sparse set of points and re­
quire markers [Uffenkamp 1993]. Computer vision approaches to
image­based modeling from calibrated photographs construct solid
object models and do not need markers [Baumgart 1974; Kutu­
lakos and Seitz 2000], but their relative accuracy is typically be­
low 1/200. [Hernandez and Schmitt 2004] propose to use the vi­
sual hull [Baumgart 1974] to initialize the deformation of a surface
mesh under the influence of rim­ and photo­consistency constraints
expressed by gradient flow forces (see [Keriven 2002] for a related
approach). Although this method yields excellent results, its re­

Source: Anderson, Richard - Department of Computer Science and Engineering, University of Washington at Seattle

Collections: Computer Technologies and Information Sciences