.nh
.po 3
.ce  
FINAL - CS528 
.nf 
Instructor Dr P Juell                   NAME_______________________ 
OPEN BOOK                                        May 20, 1992
200 points         
.fi 
1. (30)
Specify true or false:

.in +7
.ti -5
______
T(3,2) R(90)T(-3-2)
will correctly rotate an object centered at (3,2) around its center.

.ti -5
______
A true vector display driven by vector type commands does not need
a Breshenham line drawing routine.

.ti -5
______
Raytracing can produce specular highlights.

.ti -5
______
A HSV value of (0,1,1) is equal to a RGB value of (1,1,0).

.ti -5
______
You should use Gouraud shading when you want a flat surface to
appear flat.
.in -7

2. (30)
Specify the best item from the list in the blanks.
Use each item only once.

.nf
a. bilinear patch
b. Bezier curve
c. polyhedral mesh
d. Phong shading
e. surface of revolution
f. third order polynomial

_____  Estimates a surface by interpolating surface normals.

_____  Estimates a surface by cutting it up into planer polygons

_____  At least one axis can be expressed in parametric form.

_____  Estimates a curve by averaging all the points specifying the
       curve.  More weight to points closer to current location.

_____  Each point is part of a ruled surface because of two
       distinct lines.

3. (10)
You are reading the specification for a CRT. The specifications
say it is a black and white display.
The frame buffer has 2 bits per pixel with lookup table entries of 8 bits.
What does this tell you about the display.

.bp
4. (40)
You are working with a set of data that consists entirely of cylinders.
Describe, in some detail the steps you would use to produce a reasonably
shaded image.
The internal information specifies the coordinate of the center of
the bottom disk, the coordinate of the center of the top disk and the radius.
Do not use ray tracing.
Be sure to
change the standard techniques to capitalize on the special information
about the objects in the image.

.bp
5. (10)
Give a circumstance where you would prefer halftoning (super pixels) over
dithering.

.sp 5
Give a circumstance where you would prefer dithering over halftoning.
.sp 5

6. (30)
A quadtree is the 2-D version of the 3-d Octree.
Draw a quadtree for the following.  
The image is a filled circle in a square.
Show your tree, and your labeled image.
Build the quadtree to depth 2 (where needed).
Number the quads as follows:

.nf
______
| 0 1 |
| 3 2 |
-------
.fi

.bp
7. (50)
.br
a. (10)
.ul 1
Specify how to improve efficiency using techniques from the book.
Simple animations are needed and it is decided to use a program like you
developed for your last programming assignment.
That is the routine used to solid render faces using the
z-buffer technique for hidden face removal.
In addition, the program will support hierarchical segments.
Each segment will have its own transformation matrix.
The rendering process creates a copy of the segment with all points multiplied
by the segment's transformation matrix.  If the segment contains other segments,
they will be multiplied by the appropriate combination of their transformation
matrix and the enclosing segment's matrix.

For each rendering/display request, the segments 
The animation will be created by rendering sequences of images.
The motion for the animation will be specified by specifying 
new transformation matrixes for the the changed segments.
It is found that this process is extremely slow.
It is essential to speed up the operations.
As much memory and disk space as needed is available.
Specify changes that could be made by selecting other algorithms that
have been presented. 

.sp 5
b. (30)
.ul 2
Design a way to improve efficiency based on not repeating calculations
for unchanging portions of the image. 
Specify other things you could do to get the speed increases.
These are changes based on the problem, but not given in the book.
They are to use the coherence of the image from image to image.
That is much of the image will not change from image to image, we
should be able to reuse these portions of the image, if we can identify
these portions of the image.
Be specific about your algorithm.

.sp 20
c. (10)
.ul 1
Evaluate above for problems in adding shadows.
How will your additions in (b) be effected by adding shadows?
(assume the use of the z-buffer being used once for the light source
to see if the light reaches the point and then render the image.)