The USC Andrew and Erna Viterbi School of Engineering USC Signal and Image Processing Institute USC Ming Hsieh Department of Electrical Engineering University of Southern California

Technical Report USC-SIPI-159

“A Robust Algorithm for Inferring Shape from Shading”

by Qinfen Zheng and Rama Chellappa

July 1990

A robust algorithm for inferring shape from shading in a single image is presented. Assuming uniform albedo and a Lambertian surface for the imaging model, we first present methods for the estimation of illumination direction and surface albedo. Two methods for estimating the tilt of the illuminant are presented. One is based on local estimates on a smooth patch. The other method uses shading information along image contours. Both the slant angle of the illuminant and surface albedo are estimated from image statistics, taking into consideration the effects of self shadowing. The estimated parameters are then used in a shape from shading algorithm that enforces an integrability constrain, and incorporates a linear approximation for the reflectance map. Detailed experimental results on synthetic and real images are given to illustrate the usefulness of the method.

To download the report in PDF format click here: USC-SIPI-159.pdf (10.8Mb)