Computer Science 455
Instructor: R. P. Burton
Second Quiz
September 27-28, 2004
Name _________________________________________ Score ____________/52
1. Suppose you wish to generate the intersection of all scan lines with all edges for the scan-line fill algorithm. This should be done by
a. stepping along each edge with a conventional DDA algorithm
b. stepping along each edge with a conventional Bresenham algorithm
c. stepping along each edge with a DDA algorithm that always increments (or decrements) y
d. stepping along each edge with a Bresenham algorithm that always increments (or decrements) x
e. using traditional analytic geometry techniques for finding the intersection of two lines.
(c)
2. Suppose you have a polygon with vertices A, B, C, D, E, and F, positioned as shown below:
A
C
B
D
E F
Assuming the scan-line fill proceeds from the top toward the bottom, which edge should be third on the active edge list for the scan-line fill algorithm?
a. AB
b. BC
c. CD
d. DE
e. EF
f. FA
(b)
3. In the preceding example, how many scan lines intersect edge FA?
a. yA - yF
b. xA - xF
c. max (|yA - yF|, |xA - xF|)
d. it depends on the slope
(a)
4. Given "top" and "bottom" pointers to the edges in the active edge list, which of the following statements is most comprehensive, and still true?
a. All of the edges between the two pointers are still active
b. All of the active edges are between the two pointers.
c. Both (a) and (b)
d. Neither (a) nor (b)
(b)
5. For purpose of boundary fill or flood fill, a point interior to a polygon generally is specified by
a. testing candidate points with an "edge crossing" test until an interior point is found.
b. testing candidate points with a "winding number" test until an interior point is found.
c. finding a convex vertex and taking a dot product
d. finding a convex vertex and taking a cross product
e. clicking with a mouse
(e)
6. If the fill-pattern is "4 neighbors," the boundary should be ___-connected. If the fill pattern is "8 neighbors," the boundary should be ____-connected.
a. 4, 4
b. 4, 8
c. 8, 4
d. 8, 8
(c)
7. Suppose the flood-fill algorithm puts neighboring scan-lines on the stack. How does it deal with holes on the polygon or non-vertical edges?
a. it can't
b. it continually looks above (or below) as it processes the scan line
c. it looks above (or below) only when it encounters an intersection of the current scan line with a polygon boundary
(b)
8. What's wrong with using the implicit equation for a circle to generate points along its boundary?
a. nothing
b. it's slow
c. it produces unevenly spaced points
d. both (b) and (c)
(d)
9. What's wrong with using the parametric equation for a circle (i.e. stepping theta around the circumference) to generate points along its boundary?
a. nothing
b. it's slow
c. it produces unevenly spaced points
d. both (b) and (c)
(b)
10. Generally, ____-way symmetry is exploited to yield the points around the circumference of a circle.
a. 2
b. 4
c. 8
d. 16
e. infinite (well, up to the number of pixels on the circumference, anyway)
(c)
11. The "preferred" deficiency of a circle-generating DDA is that it generates
a. a circle with a tiny gap between the start and end point.
b. an ellipse rather than a circle.
(b)
12. The initial decision variable for the Midpoint Circle algorithm has a value of 5/4 - r where r is the radius of the intended circle. What does 5/4 represent?
a. The aspect ratio
b. The inverse of the aspect ratio
c. The results of a derivation with no obvious intuitive meaning
(c)
13. The midpoint ellipse algorithm generates points along the perimeter of the ellipse by stepping first in x and then stepping in y. When does the change (from stepping in x to stepping in y) take place?
a. the slope of the line from the midpoint of the ellipse to a point along the perimeter drops to 1.
b. dy/dx = -1.0
c. either (a) or (b)
(b)
14. When the coordinates of a pixel refer to the _________ , graphics primitive more accurately represent their mathematical counterparts.
a. pixel centers
b. corners of pixel "bounding boxes"
(b)
15. As a rule of thumb for drawing 2D shapes (whether in outline form or in filled form), boundary pixels should lie as nearly as possible_____
a. inside the boundary
b. on the boundary
c. outside the boundary
(a)
16. For the purpose of doing inside-outside testing by counting boundary crossings, vertices which are neither minima nor maxima should be counted as ______ crossing(s).
a. zero
b. one
c. two
(b)
17. A point is "outside" a polygon if the number of crossings from the left is ____ the number of crossings from the right, when the point is extended beyond the borders of the polygon.
a. the same as
b. different from
(a)
18. A point determined to be "inside" using the "count boundary crossings" method will also be "inside" using the "winding number" approach.
a. always
b. sometimes, but not always
c. never
(b)
19. The objective of cell encoding is
a. facilitate "inside-outside" testing
b. find intersections of self-intersecting polygons
c. generally the same as the objective of run length encoding.
(c)
20. The "not necessarily normalized" normal of the plane Ax + By + Cz + D = 0 can be determined directly from the equation of the plane. This normal is _____ the normal of the plane -Ax - By - Cz - D = 0.
a. the same as
b. different from
(b)
21. Independent of which equation (from the previous question) is used, a determination of a point's position relative to the plane produces the same results.
a. true
b. false
(b)
22. Suppose you have files which contain 1) the coordinates and "color" or each vertex, 2) the vertex pairs which constitute each edge, and 3) the vertices which constitute each face of a polyhedron. When the polyhedron is transformed (translated, scaled, rotated, sheared, reflected), how many of these three files change?
a. just one
b. two
c. all three
d. it depends on which of the listed transformations is done
(a)
23. Given the coordinates of three vertices, the equation of a plane typically is determined by
a. placing the coordinate values in their proper places in a matrix, finding the value of the determinant, and grouping.
b. finding the values of four distinct determinants
c. plugging them into four equations without any evaluation of determinants
(c)
24. In OpenGL, a vertex (i.e. a geometric entity with no dimensions) has a color.
a. true
b. false
(a)
25. Lines, as presented on graphic monitors, can have all of the following attributes except
a. width
b. color
c. texture
d. (no exceptions in this list)
(d)
26. A pixel mask such as "11110000" can be used without modification to produce uniformly textured lines.
a. true
b. false
(b)
27. Whenever a line is increased to twice its original thickness, it must shift either slightly upward or slightly downward. Hence, lines should be increased to odd multiples of their thickness.
a. true
b. false
(b)
28. Features such as butt caps, round caps, and projecting square caps are applicable to line segments; there is no similar collection of features for polyline joints.
a. true
b. false
(b)
29. If a square pixel mask measuring 15 x 15 pixels is used to plot strokes, then most points along the stroke are written to the frame buffer 14 times, especially near the middle of the stroke.
a. true
b. false
(b)
30. Suppose a 4096 x 4096 display uses a color table with 512 ( x 24 bits) entries. What percentage savings is there in the memory required for the frame buffer? Pick the range in which the answer lies.
a. 0-10%
b. 11-30%
c. 31-70%
d. 71-90%
e. 91-99.999%
(c - 67%)
31. What is "soft fill?"
a. Treating pixels as n x n (n >1) rather than 1 x 1 to increase efficiency.
b. Antialiasing to remove the "jaggies"
c. Blending (foreground and background, for example)
d. a synonym for "flood fill"
(c)
32. Virtually all text attributes have been automated with the exception of kerning, which remains an art.
a. true
b. false
(b)
33. What is the purpose of an "inquiry" function?
a. to learn what settings were used to achieve a particularly pleasing effect
b. to facilitate temporary changes
c. to verify that a change in settings has taken place
(b)
34. What is "aliasing?"
a. creating a copy of an object
b. creating a second reference to an object
c. distortion of information due to low frequency sampling
d. getting rid of the stair-step effect
(c)
35. Supersampling can be referred to as ______ ; area sampling can be referred to as _______.
a. prefiltering, prefiltering
b. prefiltering, postfiltering
c. postfiltering, prefiltering
d. postfiltering, postfiltering
(c)
36. Antialiasing tends to ______ the number of pixels used to plot a line.
a. decrease
b. leave unchanged
c. increase
(c)
37. Pixel-weighted masks have the same effect as filtering techniques.
a. true; they are synonyms
b. false; pixel-weighted masks integrate over the pixel surface
c. false; filtering techniques integrate over the pixel surface
(c)
38. Pixel phasing involves
a. supersampling
b. area sampling
c. either pixel-weighted masks or filtering techniques
d. micropositioning of the electron beam
(d)
39. The boundary between a blue polygon and a red polygon can be made to appear straighter by imposing a line of various shades of magenta along the boundary.
a. true
b. false
(a)
40. The magnitude of the decision variable from the Bresenham algorithm can be used as a reasonable estimate of pixel overlap.
a. true
b. false (Wouldn't it be nice if the magnitude, not just the sign of the decision variable had meaning.)
(a)
41. An object subjected to translation may deform, depending on the shape and/or the orientation of the object prior to the translation.
a. true
b. false
(b)
42. When a two-dimensional object is rotated using a rotation matrix with 2 sines, 2 cosines, and 1s otherwise on the diagonal, it is rotated about
a. its upper left corner
b. its lower right corner
c. its "center of gravity"
d. the origin
(d)
43. Scaling can affect
a. the dimensions of an object
b. the position of an object
c. both (a) and (b)
(c)
44. A point in homogeneous coordinate space corresponds to _______ point(s) in the space of dimension less by one.
a. one
b. two
c. infinitely many
(a)
45. Homogeneous coordinates are needed to accommodate ______ in composable matrix transformation form.
a. translation
b. rotation
c. scaling
d. rotation and scaling
e. all of the above
(a)
46. Suppose you have a transformation matrix which rotates an object about a point which is not the origin. If the execution could be slowed way down, and the object were plotted at each stage of execution, you would be able to observe that the rotation point moves to the origin, the object is rotated, and the rotation point moves back to its original position.
a. true
b. false
(b)
47. For the purposes of transformations, matrix multiplication in general is
a. associative
b. commutative
c. both (a) and (b)
d. neither (a) nor (b)
(a)
48. The inverse of a translation matrix may be obtained by
a. complementing the translation elements
b. inverting the translation elements
c. transposing the translation elements
d. none of the above
(a)
49. The inverse of a scaling matrix may be obtained by
a. complementing the scaling elements
b. inverting the scaling elements
c. transposing the scaling elements
d. none of the above
(b)
50. The inverse of a composite transformation matrix may be obtained by
a. complementing all the elements
b. inverting all the elements
c. transposing all the elements
d. none of the above
(d)
51. The y-axis points left, the x-axis points down. For counterclockwise rotation, the sine with the minus sign should be on the ______ row.
a. first
b. second
(b)
52. For n-dimensional translation, the number of nonzero elements in the transformation matrix is
a. n
b. 2n -1
c. 2n
d. 2n + 1
e. none of the above
(d)