Statistical Methods for Surface Integration
In this paper we show how statistical constraints can be incorporated into the surface integration process. This problem aims to reconstruct the surface height function from a noisy field of surface normals. We propose two methods that employ a statistical model that captures variations in surface height. The first uses a coupled model that captures the variation in a training set of face surfaces in both the surface normal and surface height domain. The second is based on finding the parameters of a surface height model directly from a field of surface normals. We present experiments on ground truth face data and compare the results of the two methods with an existing surface integration technique.
KeywordsRoot Mean Square Error Root Mean Square Couple Model Surface Normal Surface Integration
Unable to display preview. Download preview PDF.
- 5.Wei, T., Klette, R.: Height from gradient using surface curvature and area constraints. In: Proc. ICVGIP, pp. 204–210 (2002)Google Scholar
- 6.Agrawal, A., Chellappa, R., Raskar, R.: An algebraic approach to surface reconstruction from gradient fields. In: Proc. ICCV 1, 174–181 (2005)Google Scholar
- 8.Cootes, T.F., Taylor, C.J.: Statistical models of appearance for computer vision. Technical report, Division of Imaging Science and Biomedical Engineering, University of Manchester (2001)Google Scholar
- 9.USF HumanID 3D Face Database, Courtesy of Sudeep. Sarkar, University of South Florida, Tampa, FLGoogle Scholar
- 10.Smith, W.A.P., Hancock, E.R.: Face recognition using 2.5d shape information. Proc. CVPR 2, 1407–1414 (2006)Google Scholar