Gross - 2 
1 Graphics Models for Vision 
In the past decades, two research lines developed almost independently from each other: 
The one, computer graphics, deals with the generation of images from computer-based 
models and with the problem of describing these models efficiently using advanced calcu 
lus. The other, computer vision, treats the inverse process of interpreting pixel-based 
images, of extracting symbolic information and of recovering the underlying geometry. 
Obviously, both disciplines are complementary parts of one common understanding and 
they allow to trace the path from geometry to the image and back again. Later [1], in 1987, 
scientific visualization was bom and defined as interdisciplinary research, comprising 
methods from different fields. 
Moreover, due to the increasing complexity of many advanced applications, scientists 
from both areas began to take over some of the sophisticated techniques developed in the 
past and formed a new research line: Visual Computing [5]. This newly created area targets 
at promoting the interaction between graphics and vision, as illustrated in fig. 1. 
Computer 
Vision 
Modeling: 
-geometry 
-topology 
-physics 
Fig. 1 Interactions between computer graphics and computer vision and the key role of 
models. 
It is clear, that there are many topics well suited for an essay on how graphics and vision 
might fertilize each other, but one of the most striking ones are models. Models can be con 
sidered in many different ways, depending on what the tasks are they have been tailored 
for. Generally, in graphics we keep concentrated on either: 
• Pure geometric models. 
• Topology and advanced data structures. 
• Physically-based models.