Spring Quarter: Tues, Thurs 4:30-6:00pm
Instructor: Jack Tumblin ( jet@cs.northwestern.edu
, www.cs.northwestern.edu/~jet
)
Recent computer graphics work promotes digital images to a
'first class' primitive, to make images equally useful as both input and
output. IBMR methods use images in new ways to address questions like
these:
Images hold partial information about lighting, objects, surfaces, positions and movements--that's why people like to look at them. This course will introduce you to some recent techniques that try to extract, modify, and/or re-use some of this information to make new images.
Prerequisites--CS-351 Introduction to Computer Graphics,
reasonable comfort with linear algebra
or permission of instructor (please ask if you're
interested--you can probably do it!)
Topics: projective image geometry and warping, image resampling and antialiasing, warping images with depth, epipolar geometry, recovering geometry, 'inverse rendering' problems, high-contrast imaging, BRDF recovery.
Text:
Multiple-View Geometry, Hartley and
Zisserman, Cambridge Press, plus handouts.
This is a good, comprehensive text, though it looks more intimidating than it
actually is.
Plan: We will closely (and slowly) examine Part I and Part II of the text (Camera Calibration, Two-View Geometry/Epipolar Methods). Unlike winter quarter's 'Advanced Graphics Seminar', this course will be narrower, will require programming in Visual C++. You will learn strengths and weaknesses of methods by actually trying them. We may use 100-year old stereo-opticon slides from the Eastman Kodak museum, famous paintings, computer graphics renderings, or digital photos taken in class or in the lab.
Grading: 1 midterm, no final, and one programming project built and graded in 5 stages. Projects will receive "Image-based grading"--let's make some interesting pictures!