Two-Dimensional

Two Dimensional
Viewing

Chapter 6
Intro. to Computer Graphics
Spring 2008, Y. G. Shin
Viewing
e g Pipeline
pe e
Two-Dimensional
o e s o a Viewing
e g
„ Two dimensional viewing transformation
„ From world coordinate scene description to
d i (screen)
device ( ) coordinates
di t
Normalization and Viewport
Transformation
„ World coordinate clipping window
„ Normalization square: usually [-1,1]x[-1,1]
„ D i coordinate
Device di t viewport
i t
 C pp g
Clipping
ƒ Remove portion of line outside viewport or
screen boundaries
ƒ Two approaches:
ƒ Clip during scan conversion: per-pixel
per pixel bounds check,
or span endpoint tests.
ƒ Clip analytically,
analytically then scan-convert the modified
primitive.
Two-Dimensional
o e so a C Clipping
pp g
„ Point clipping – trivial
„ Line clipping
„ Cohen-Sutherland
Cohen Sutherland
„ Cyrus-beck
„ Liang Barsky
Liang-Barsky
„ Fill-area clipping
„ Sutherland-Hodgeman
S th l d H d
„ Weiler-Atherton
„ Curve clipping
C li i
„ Text clipping
 Line
eCClipping
pp g
„ Basic calculations:
„ Is an endpoint inside or outside the clip rectangle?
„ Fi d the
Find th point
i t off intersection,
i t ti if any, b
between
t a liline
segment and an edge of the clip rectangle.
9Both endpoints inside:
trivial accept
9One inside: find
p
intersection and clip
9Both outside: either
clip or reject
Cohen-Sutherland
Line-Clipping
Line Clipping Algorithm

1001 1000 1010

View port

0001 0000 0010

0101 0100 0110

< Region code for each endpoint >

above below right left
Bit 4 3 2 1
Cohen-Sutherland
Line-Clipping
Line Clipping Algorithm
„ Trivially accepted
if (both region codes = 0000)
„ Trivially
ll rejected
d
if ((AND of region
g codes ≠ 0000))
„ Otherwise, divide line into two segments
„ test intersection edges in a fixed order.
order
(e.g., top-to-bottom, right-to-left)
Cohen-Sutherland
Line-Clipping
Line Clipping Algorithm
* fixed order testing and clipping cause
needless clipping (external intersection)
Cohen-Sutherland
Line-Clipping
Line Clipping Algorithm
„ Midpoint Subdivision for locating
intersections
1. trivial accept/reject test
2. midpoint subdivision:
xm = (x1 + x2)/2, ym = (y1 + y2)/2
(one addition and one shift)
3. repeat step 1 with two halves of line
⇒ good for hardware implementation
Cohen-Sutherland
Line-Clipping
Line Clipping Algorithm
„ When this is good
„ If it can trivially reject most cases
„ Works well if a window is large w.r.t. to data
„ Works
o s well
e if a windowdo iss small
s a w.r.t.
t to data
„ i.e., it works well in extreme cases
„ Good for hardware implementation
Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)
„ Use a parametric line equation
P(t ) = P0 + t ( P1 − P0 ), 0 ≤ t ≤ 1

„ Reduce
educe tthe
e number
u be oof ca
calculating
cu at g
intersections by simple comparisons of
parameter t.
Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)
Algorithm
g
„ For each edge Ei of the clip region
„ N i : outward normal of Ei
 Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)
„ Choose an arbitrary point PEi on edge Ei and
consider three vectors P(t ) − PEi
⇒
N i • ( P (t ) − PEi ) < 0 ⇔ a point in the side halfplane
N i • ( P (t ) − PEi ) = 0 ⇔ a point on the line containing the edge
N i • ( P (t ) − PEi ) > 0 ⇔ a point in the outside halfplane
Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)
„ Solve for the value of t at the intersection
of P0P1 with the edge:
Ni · [P(t) - PEi] = 0.
P(t)
( ) = P0 + t(P
( 1 - P0) and let D = ((P1 - P0),
Then
N i ⋅ [ P0 − PEi ]
t=
− Ni ⋅ D
Ni ≠ 0,
0
D ≠ 0 (that is P0 ≠ P1),
Ni · D ≠ 0 (if not, no intersection)
Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)
„ Given the four values of t for a line segment,
segment
determine which pair of t's are internal
intersections.
If t ∉ [0,1] then discard
else choose a (PE,
(PE PL) pair
that defines the
clipped line.
line

„ PE(potentially
(p y entering)
g) intersection:
if moving from P0 to P1 causes us to cross an edge to
enter the edge
edge'ss inside half plane;
Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)
„ PL(potentially leaving) intersection:
„ if moving from P0 to P1 causes us to leave the
edge's
g inside half plane.
p

i.e., N i • P0 P1 < 0 ⇒ PE
N i • P0 P1 > 0 ⇒ PL

„ Intersections can be categorized!

„ Inside the clipp rectangleg (T( E,,TL)
„ TE: select PE with largest t value ≥ 0

„ TL: select PL with the smallest t value ≤ 1.

1
 Parametric Line Clipping
(Cyrus-beck
(Cyrus beck Technique)

„ This is
Thi i an efficient
ffi i t algorithm
l ith when h many line
li
segments need to be clipped
„ C be
Can b extended
d d easily
il to convex polygon
l windows
i d
Liang-Barsky
a g a s y line
e clipping
pp g
„ The ideas for clipping line of Liang
Liang-Barsky
Barsky and
Cyrus-Beck are the same. The only difference is
Liang-Barsky
Liang Barsky algorithm has been optimized for
an upright rectangular clip window.
„ Finds the appropriate end points with more
efficient computations.
 Liang-Barsky
a g a s y line
e clipping
pp g
Let PQ
Q be the line
which we want to study

Q(x2,y2) t Parametric equation of

R
the line segment
tT
x = x1 + ( x 2 − x1)t = x1 + dx × t
P(x1 y1)
P(x1,y1) tL y = y1 + ( y 2 − y1)t = y1 + dy
d ×t

tB
t = 0 ⇒ P( x1, y1)
t = 1 ⇒ Q( x 2, y 2)
 Liang-Barsky
Liang Barsky Line Clipping
1. Set tmin = 0 and tmax = 1
2. Calculate the values of tT, tB, tL, tR,

L R
tT T

B
tB
 Liang-Barsky
Liang Barsky Line Clipping
„ If t < tmin or t > tmax, ignore
g it and go
g to the
next edge. N
„ Otherwise classify the t value as entering or
exiting value (using the inner product to tT
classify)
„ Let PQ be the line and N is normal vector tR
„ If N • (Q − P ) ≤ 0, the parameter t is entering
„ If N • (Q − P) > 0 , the parameter t is exiting
„ If t is entering value, set tmin = t, if t is exiting
value set tmax = t
Liang-Barsky
a g a s y Line
eCClipping
pp g
3 If tmin < tmax then draw a line
3.
from ( x1 + dx × t min , y1 + dy × t min )
to ( x1 + dx
d × t max , y1 + dy
d × t max )

Q(x2,y2) t
R

tT

P(x1,y1) tL

tB
 C pp g
Clipping
„ Clipping rotated windows,
windows circles
„ trivial acceptance/rejection test with respect to
bo nding rectangle
bounding ectangle of the window
indo
„ Line clipping using nonrectangular clip
windows
„ extend Cyrus
Cyrus-Beck
Beck algorithm
Polygon
o ygo clipping
pp g
„ Sutherland-Hodgeman
Sutherland Hodgeman Algorithm
„ clip against 4 infinite clip edge in succession
 Sutherland-Hodgeman
Sutherland Hodgeman Algorithm
„ Accept a series of vertices (polygon) and outputs
another series of vertices
„ Fo possible outputs
Four o tp ts
 Sutherland-Hodgeman
Sutherland Hodgeman Algorithm
„ The algorithm correctly clips convex polygons,
polygons
but may display extraneous lines for concave
polygons.
polygons

„ How clip?
How
o to
o correctly
o e y clip
p
[Way I] Split the concave polygon into two
or more convex polygons and process each
convex polygon separately.
[Way II] Modify the algorithm to check the
final vertex list for multiple vertex points
along any clip window boundary and
correctly join pairs of vertices.
[Way III] Use a more general polygon
clipper
Clipping
C pp g concave
o a e polygons
po ygo s
„ Split
p the concave polygon
p yg into two or more
convex polygons and process each convex
polygon separately.
separately
„ vector method for splitting concave polygons
⇒ calculate edge-vector
edge vector cross products in a
counterclockwise order. If any z component turns out to
be negative, the polygon is concave.
Weiler-Atherton
Weiler Atherton Polygon Clipping
„ For an outside-to-inside p
pair of vertices,, follow the
polygon boundary.
„ For an inside
inside-to-outside
to outside pair of vertices, follow the
window boundary in a clockwise direction.
Weiler-Atherton
Weiler Atherton Polygon Clipping
„ Polygon clipping using nonrectangular
polygon clip windows
 Texture Clipping
1. all-or-none text clipping
pp g : Using
g boundaryy box
for the entire text
2. all-or-none character clipping : Using boundary
box for each individual
3. clip individual characters
„ vector : clip line segments
„ bitmap : clip individual pixels
What we have got!
g