Learning a near-optimal estimator for surface shape from shading

David C. Knill, Daniel Kersten

Research output: Contribution to journalArticlepeer-review

7 Scopus citations

Abstract

The problem of determining surface shape from shading is formulated in terms of Bayesian estimation. The goal is to select an estimate of surface shape that best fits some criterion on the posterior probability of the surface conditional on the image data. This conditional probability is a function of the imaging function and the prior probability of the surface. A gradient descent technique is used to compute the best linear estimator of the mean of the conditional distribution from a set of random fractal surfaces and their images. Simulations show that the derived estimator works well for a wide range of surfaces, including non-fractals. Once learned, the estimator could be used to implement a fast, non-iterative parallel device for estimating shape from shading with arbitrary light source directions. The estimator has similar perceptual deficits to human observers. In particular, it exhibits a loss of coarse-scale surface shape, while accurately reconstructing fine-scale details.

Original languageEnglish (US)
Pages (from-to)75-100
Number of pages26
JournalComputer Vision, Graphics and Image Processing
Volume50
Issue number1
DOIs
StatePublished - Apr 1990

Bibliographical note

Funding Information:
*This research was supported by BRSG PHB 507 RR 07085 and NSF Grant BNS-8708532 to Daniel Kersten, and by NSF BNS-85-18675 and ONR N-00014-86-K-0600 to James A. Anderson. The authors would like to acknowledge the help and support of the neural mode&g group at Brown University. We would especially like to thank Jim Anderson and Mike Rossen for many helpful discussions and suggestions. iPresent address: Department of Psychology, University of Minnesota, Minneapolis, MN 55455.

Fingerprint Dive into the research topics of 'Learning a near-optimal estimator for surface shape from shading'. Together they form a unique fingerprint.

Cite this