Feb 26, 2008

1. What does "post-multiply" mean in the stage1 of project 2?

It's the same convention we used in class and OpenGL. If you say

glmatrix(M);

glmatrix(N);

then you will get a matrix M*N. What "post" means is that N is multiplied after M.

2. In question 7, what is the order of "composing a rotation and a uniform scale"?

scale*rotation

Let g = 1/5 [1 1 1 1 1]. This can mean different things.

One way to interpret it is: g(1)=g(2)=g(3)=g(4)=g(5)=1/5 and g is zero everywhere else.

By “zero centered”, it means g(-2)=g(-1)=g(0)=g(1)=g(2) = 1/5 and g is zeros everywhere else.

2. Correction of question 2 in the written assignment (stage 3) of project 1

WAS: 2.A The unit box (g(t) = 1 if -.5 < t <= .5, 0 otherwise)
NEW: 2.A The unit box (g(t) = 1 if -.5 <= t < .5, 0 otherwise)

WAS: 2.B ... Note: in the book, this is the tent filter of r=1/2
NEW: 2.B ... Note: in the book, this is the tent filter of r=1

WAS: 2.D The tent of r=1
NEW: 2.D The tent of r=2

For the command "chromakey" in project 1, you can assume that initially V has alpha = 255 for every pixel.

For the "fill" command, if the input region includes pixels that are outside the image, your code should fill in the intersection of the image and the input region. If the region is invalid, i.e., y2<y1 or x2<x1, you code does not need to do any job -- just leave the image as it is.

For the two tutorials on alpha compositing, Alvy Ray Smith, "Image Compositing Fundamentals", page 1-6 till the end of "premultiplication problems", is required; other sections are optional. Tom Porter and Tom Duff  "Compositing Digital Images" is also optional.