Shape estimation from topographic primal sketch |
| |
Authors: | Ting-Chuen Pong Linda G. Shapiro Robert M. Haralick |
| |
Affiliation: | Computer Science Department, University of Minnesota, Minneapolis, MN 55455, U.S.A.;Machine Vision International, 325 E. Eisenhower, Ann Arbor, MI 48104, U.S.A. |
| |
Abstract: | ![]() In this paper, a technique for estimating object shape from topographic primal sketch is investigated. Given a gray tone image of a three-dimensional object, a topographic labeling of the image indicates the peaks and pits, ridges and valleys, and flats and hillsides of the underlying, continuous, gray tone surface. The patterns of these topographic labels capture information about the original three-dimensional object in the scene and about the illumination. In order to determine if estimation of three-dimensional shape from a topographic labeling is feasible, we have both analytically and experimentally determined the topographic labelings for images of some mathematically generated surfaces with varied directions of illumination. Our results indicate that such patterns do exist and will be useful in determining three-dimensional shape from two-dimensional images. A scheme for partial classification of three-dimensional object surface is proposed. Preliminary results are illustrated. |
| |
Keywords: | Topographic primal sketch Computer vision Three-dimensional shape Surface classification Topographic patterns |
本文献已被 ScienceDirect 等数据库收录! |
|