INSTRUCTOR:
Prof. Andrew Knyazev
Office: CU (Dravo Bldg) 644. Phone: (303) 556-8102.
Office hours: by appointment.
WWW: http://www-math.cudenver.edu/~aknyazev
TEXTBOOK:
Approximation Theory and Methods,
M. J. J. D. Powell.
Format:Textbook Paperback, 1st
ed., 352pp.
ISBN: 0521295149
Publisher:
Cambridge University Press.
Pub. Date: October 1981
Midterm project: Comparing Apples and Oranges with
MATLAB Image Processing Toolbox
-
Learn the basics of the MATLAB
Image Processing Toolbox, e.g.,
how to use the
imread function.
-
Run the Image Processing Toolbox Demo called
"Identifying Round Objects" in the
"Measuring Image Features" Demo chapter
and try to understand what it does.
-
Read into MATLAB provided images of
apples and
oranges.
-
Find a formula for image comparison
that would give a small distance
between two selected (or preprocessed, see below)
images of an apple and
a large distance between images of an apple and
an orange.
-
Using this distance formula, find the best
approximation of this simple image to (a) the
properly chosen subset of
images of apples and (o) the properly chosen subset of
images of oranges. Image Identification:
Based on the approximation errors,
conclude if the given image is of an apple or an orange.
-
For the advanced project, preprocess all
advanced analysis images
of apples and oranges using the analog of the
"Identifying Round Objects" Demo finding the actual
locations and the sizes of apples and oranges,
and perform the cut and scaling prior to
applying the distance formula and finding the
best approximations. Perform the Image Identification
using the approximation errors.
A sample project report
by
Eugene Vecharynski.