Assignments: Sensory Motor Systems, CSC 449 (Fall 2002)
Science is a game --- but a game with reality, a game played with
sharpened knives... if a man cuts a picture carefully into 1,000
pieces, you solve the puzzle when you reassemble the pieces into a
picture; in the succcess or failure, both your intelligences compete.
In the presentation of a scientific problem, the other player is the
good Lord. He has not only set the problem but also devised the rules
of the game --- but they are not completely known, half of them are
left for you to discover or to determine. The experiment is the
temmpered blade which you wield with success against the spirits of
darkness -- or which defeats you shamefully. The uncertainty is how
much of the rules God himself has permanently ordained, and how much
appears to be caused by your mental inertia, while the solution
generally becomes possible only through freedom from this limitation.
This is perhaps the most exciting thing in the game. For here you
strive against the imaginary boundary between yourrself and the
Godhead -- a boundary that perhaps does not exist.
Making Digital Images
Simulation and Calibration
Clustering and Classification
The subject is up to you. First you should write and turn
in a proposal justifying why the topic is interesting and doable,
what you propose to accomplish. PLEASE feel free to bring your ideas
to the Prof or TA (just a few words after class can be very valuable)
before writing your proposal.
It can be strictly a written
(scholarly, researchy) paper -- my guideline would be "20 references
(at least half from books and refereed papers)
connected by a good story". You can do your own research
(mathematical, psychophysical, programming), you can write a
critical review of
the state of some area that interests you (The Turing Test: its
relevance and how close have we come?;
The Promise and Disappointments of Neural Nets,
The Possibilities and Difficulties of
Understanding Video, Successes and Failures of Automatic Planning),
Or it can be a programming project. Several quite good ones
appear in the problem sections of your text, with little keyboards
next to them. Or make up your own.
Here's last year's third programming problem,
a Telling O's from X's
Last year I was not terribly
impressed by what people did, in general: The key is that it's just
not interesting to solve a simple case of character recognition. But
it is nice when you can make up your own data. So the key is to
create challenging problems within a simple format. NOT hard given
the various text fonts out there, including handwriting. So here is
a large Database of digitized
digits to play with.
One other partially-field-tested project is
Robotic Control Assignment from the
undergraduate robotics course.
A local project currently running: Grad student Matt Boutell is
working on scene classification for Kodak. Issues of high-level scene
interpretation down to low-level features, from strategy to tactics,
from learning systems to expert systems.
Another local project currently running: UG Jon Schmid is doing
mobile robot obstacle detection and navigation using optical flow. We
have had some success and have written things up more or less,
but there is more to do.
Yet another local project currently running: UG Jon Schmid
has pursued stereo input for obstacle detection but it is challenging
given the tools he had. You could implement a good stereo algorithm
and make it run on the mobile robot with his collaboration.
Some inspiration: how about this... a Virtual phonograph .
More background and possible links to related sites at this
Longwinded thread of typical geeky maunderings .
Or this famous project in Flow Control?
Mostly non-Programming Assignments
The keys EX1, EX2... correspond to the entries in the
syllabus. Some exercises may have their own links, others appear here
directly. Ch4 means Chapter 4 of Text, Sonka et al.
In general it's a good idea to go through all the exercises for
the current chapter to make sure you feel comfortable with them.
In particular, you should make sure you can answer all the "short answer" questions
for each chapter you read. Some of them actually involve a little
(e.g. find the Fourier tranform of the Dirac delta function), but
most just give you a chance to restate concepts defined in the text.
The Text's "Problems" are often of the form: Write code to implement X
and apply it to a range of images for a range of parameters.
This habit of methodically
exploring the multidimensional space of variables in a vision
problem is a good one to cultivate! But it's not too much of an
exaggeration that several of these problems from any chapter
could easily assume the scale of a term project! Thus we won't be
doing many of the book's problems.
Instead, I'll post some problems every week or so depending on what
seems fun and relevant at the time. They'll be more mathematical or
geometric, not programming.