In 1977 we proposed a computational model of visual perception which focused our own work and influenced that of a number of others (Barrow and Tenenbaum [4]). A key feature was the simultaneous recovery of retinotopic arrays of characteristics intrinsic to the scene, such as surface color, depth, a
Emulating the human interpretation of line-drawings as three-dimensional objects
β Scribed by Thomas Marill
- Publisher
- Springer US
- Year
- 1991
- Tongue
- English
- Weight
- 929 KB
- Volume
- 6
- Category
- Article
- ISSN
- 0920-5691
No coin nor oath required. For personal study only.
β¦ Synopsis
The human vision system has the ability to interpret two-dimensional images as three-dimensional objects. In this article, we present a program that emulates this ability for the case of images consisting of line-drawings. As a by-product of the approach, we provide an explanation of the Necker cube illusion.
1 Problem Definition How does the human vision system produce threedimensional interpretations of two-dimensional images? Can the process be emulated by computer? These fundamental questions form the basis of our inquiry. We do not propose, however, to deal with images in their full generality. Instead, we note that the phenomenon of three-dimensional interpretation holds for a class of very simple images, namely line-drawings ("a noble class of pictures," as Sugihara (1986) calls them), and we focus on this class.
To say that we wish to study the three-dimensional interpretation of line-drawings is not to pose a precise question, however. What exactly is the problem to be solved? There are at least three alternative problem definitions from which to choose.
(1) The inverse-optics formulation: Under this formulation, the problem is to create a program that recovers the object that produced the input image. The criterion of success of a particular program is the degree to which the output object generated by the program agrees with the object that produced the input image.
(2) The model-basedformulation:
In this formulation, the program is given a set of models and a set of transformations to apply to the models. The program's task is to select models from among the given *Support for the Artificial Intelligence Laboratory is provided in part by the Advanced Reserach Projects Agency of the Department of Defense under Office of Naval Research contract N00014-85-K-0124.
π SIMILAR VOLUMES
Drawings of curved objects often contain many linear features: straight lines, colinear or coplanar points, parallel lines and vanishing points. These linear features give rise to linear constraints on the 3D position of scene points. The resulting problem can be solved by standard linear programmin
Although the combination of cisplatin and etoposide has been used as standard therapy for small-cell lung cancer, it is difficult to demonstrate combination effects between cisplatin and etoposide in vitro. We therefore adopted a 3-dimensional (3-D) model to analyze the combination effects of antica