Clarifications and tips for Homework #1
---------------------------------------

The due date for Homework #1 has been changed.  It is now due FRIDAY, APRIL 21.

Problem 1c: 

  The answer to this could be argued as *either* true *or* false.  The
  key in answering this question is justifying your position.

Problem 1e and 1g: 

  You should be thinking of RGB and HSV as somewhat abstract color
  spaces -- i.e., ways of "naming" (or "enumerating") a perceived color.
  RGB does, of course, have a physical realization as voltages that
  drive electron guns that send electrons to phosphors that emit
  spectra.  In this case, think of RGB as corresponding not to those
  spectra, but to the perception of color that they cause.

Problem 2b: 

  The filters are to be treated "as is".  You are not to sum the
  filter weights and divide.  Just analyze the problem using the
  weights exactly as given.

Problem 2c: 

  The output image gets created from the input image by stepping through
  (x,y) coordinates of the output image, and then figuring out which
  pixels (X,Y) need to be plucked from the input image.  Another way to
  phrase this is that g(x,y) = f(X(x,y),Y(x,y)), where f is the input
  image and g is the output image.  Recall in class that things happen
  in a perhaps unexpected way when you do this.  When we said g(x,y) =
  f(2x, 2y), the output image actually *shrank* by 1/2 in x and y.

  When thinking about coordinate systems, place the origin at
  the center of the image.

  Also, we had a sign error in the program that generated the image pairs.
  We corrected the error, and we also changed one of the images (IV) a
  little to make it easier to understand what was going on.  The new
  images are in place on the homework web page in both Word and PDF
  form.  You can look at these for reference or you can print the
  pages and attach them to your homework.  In the end, we will just be
  checking to see which equation number you assigned to each Roman
  numeral image pair.

Problem 3a:

  The A, B, C spectra basically have nothing to do with the spectra
  with similar names on the 3rd and 4th pages of your color lecture
  notes -- it's just a coincidence.  You could just as easily rename
  the spectra in this problem u,v,w.

  Also, the first plot in this problem shows the sensitivities of the
  s and l cones, so there are really two functions on that plot.  All
  other spectral plots are single functions (that go to zero in
  places).