CS 348C Project Proposal
CS 348C - Modeling in Computer
Graphics
Fall quarter, 1995
Coordinator: Apostolos Lerios
The project proposal is a 1-2 page document that describes the
intended area of a student project, summarizes relevant published
work, and suggests possibilities for the concrete goal of the project.
Content
- Intended area
- Students must select an area related to geometric modeling and/or
modeling of natural phenomena and processes. Different teams may work
on the same area; however, the concrete goals of their projects have
to be different. The topics list contains
several suggestions on possible areas that a project may
investigate. Also, the staff of Pacific
Data Images has made the following suggestions:
- Meltdown; crumbling down objects.
- Earthquakes.
- Rain and snow, with an emphasis on controlling their flow.
- Constrained particle systems, such as water flowing out of a
nozzle and going down a sink.
- Behavior-guided animal movement, such as flocking and herding.
- Modeling skin and ageing.
- Light shafts and rainbows, and any other subtle natural tricks of
light.
- Formation and destruction of galaxies.
- Seasonal variation of trees and forests.
Although by no means they are required to do so, most students pick a
topic related to the papers they researched for their in-class presentations.
- Relevant work
- Students must reference published work which relates to the
concrete goal of their project. References outside the intended area,
but relevant to the concrete goal, are highly desirable. In addition
to merely listing bibliographical references, students must append a
short (up to 10 lines) content summary to each referenced
item.
- Concrete goal
- Students must propose a concrete focus of study within the
intended area. For example, the students may opt to
- implement and/or optimize a particular algorithm,
- attempt to solve a specific problem,
- compare and combine two published models,
- design and implement a new modeling technique, or
- craft a new, friendly user interface to an existing modeling
primitive.
The students are encouraged to read carefully the "future work"
sections of relevant published work for ideas. Also, it's a good idea
to contact the authors of relevant papers and ask them if they have
any particular projects to suggest.
Outline of sample proposal
[ Editorial remarks appear in square brackets. ]
[ This is just an outline. A full-length proposal should have a
similar structure, but the students should elaborate more on their
ideas. ]
- Intended area
- We would like investigate the modeling of plants.
- Relevant work
- [ Four sample references are shown; a full-length proposal
should reflect a more extensive literature search. ]
- Prusinkiewicz, Przemyslaw, and Aristid Lindenmayer. The
Algorithmic Beauty of Plants. New York: Springer, 1990. The
authors present a variety of techniques for modeling the structure of
plants. The basic technique they use is called L-systems, and
is akin to grammars used by parsers.
- Prusinkiewicz, Przemyslaw, Mark S. Hammel, and Eric
Mjolsness. Animation of Plant Development. Proceedings of
SIGGRAPH '93: 351-360. This paper extends basic L-systems, which can
only develop structure in discrete steps, to dL-systems
(differential L-systems), which can encode the continuous growth of
plants. The new formalism uses (1) grammars and rewriting rules to
model discrete transitions in a plant's growth, and (2) differential
equations to model the gradual changes a plant undergoes between
discrete transitions.
- Aho, Alfred V., Ravi Sethi, and Jeffrey
D. Ullman. Compilers, Principles, Techniques, and Tools.
Reading, MA: Addison, 1986. This is one of the definitive textbooks in
compilers. Relevant to our project is the detailed presentation of
parsing techniques, esp. the discussions on efficiency issues.
- Eldén, Lars and Linde Wittmeyer-Koch. Numerical
Analysis: An Introduction. San Diego: Academic P., 1990. This
is one of the definitive textbooks in numerical analysis. Relevant to
our project is the discussion of differential equations and solution
methods, esp. issues regarding numerical stability and convergence
speed.
- Concrete goal
- [ A full-length proposal should be more detailed than this
rough sketch. Also, several alternative goals may be presented. ]
We plan to implement a dL-system parser, which will improve upon the
work of Prusinkiewicz in three areas:
- We would like to design a friendly user interface to
dL-systems. That is, we would like to design an easy-to-learn-and-use
specification language for entering the modules of the dL-system;
these include productions, differential equation, their domains, and
their boundaries.
- We would like our dL-system parser to be efficient and
stable. That is, it should be able to robustly generate
plants at interactive speeds so that the animator can
comfortably experiment with the dL-system parameters. This way, (s)he
may gain intuition on the effect individual system parameters have on
the generated plants and thus quickly design desired plants.
- We would also like to extend dL-systems by allowing the user to
specify continuous growth by arbitrary procedures, rather than mere
differential equations. Such an extension raises several issues
including the following: what is the proper user interface to such
procedures (maybe PERL scripts)? How can the parser reliably predict
when and if modules get (de-)activated?
Malleability
The project proposal is not unlike a Ph.D. thesis proposal. Here are
the requirements of the latter, as set forth by the computer science department and
paraphrased for CS 348C:
On the one hand, this proposal is a contract. It commits the students
to begin work in the proposed area and it binds the coordinator to
advise the students in that area. On the other hand, it is just a
proposal. There is no requirement that the students continue working
in the proposed area. In fact, few students end up doing exactly what
they initially propose. The students, with the approval of the
coordinator, can change the area, goal, and scope of the project
at any time.
Last update: 13 December 1995 by Apostolos "Toli" Lerios
tolis@cs.stanford.edu