CSc Senior Capstone Sequence 2006-2007
Computer Science - The City College of New York

Computational Imaging

Instructor: Michael Grossberg

Inexpensive and high quality imagers are now everywhere. Above deep in
space satellites send us streams of images with which we can monitor
and explore earth and is neighboring planets. In the deepest recesses
of the ocean, robots bring us images of an environment where life
thrives in an alien environment under extreme pressure and
temperature. New imaging technology allows us to probe the structure
and function of our brains. Cameras now make video conferencing
between distant individuals commonplace. Imaging devices are also
enhancing everyday technology, giving a car the ability to sense
dangerous road conditions, and a robotic vacuum the ability to
navigate around a room.  With a vast quantity of rich imaging data we
are faced with the dilemma that it is impossible for humans to examine
more than a tiny fraction of these images. Fortunately, we now have
the computing power and fundamental algorithms to begin to explore
this wealth of visual information, automatically.

In this two semester capstone course we will investigate how we can
automatically answer questions about the world using a set images from
one or more sensors. The focus will be on problems with a large number
of degrees of freedom (high dimension) and many images. In order to
deal with this we will use physical models wherever possible to reduce
the complexity, and data mining/machine learning techniques for
automatic analysis.

Topics may include:

    * Models of Imaging
    * Calibration
    * Physics of light and reflectance
    * Image feature extraction such as edges, corners
    * Video feature extraction, such as optical flow
    * Dimension reduction
    * Techniques from  machine learning such as  clustering and classification

Projects currently under consideration:

    * Compression of images of the earth from the GOES and Metrosat satellites
    * Classification of pathologies in diffusion Tensor MRIs of the brain
    * Capturing the appearance a large urban scene from photograps
    * Robot navigation and localization from video

The projects are subject to changes based on the first semester course
progress. Groups of students may work in teams on the same project.
In the second semester, the project teams formed in the first semester
will mainly focus on the implementations of the projects.
Implementation languages and libraries used will be selected by the
project teams.

You can contact Michael Grossberg, at email address grossberg at cs
dot ccny dot cuny dot edu.