-553

CS-553 Interactive Computer Graphics



Lecture 2:
Linear Algebra for CG

G. Papagiannakis

Interactive Computer Graphics, Spring 2013, University of Crete



Overview

a

Basic

linear algebra
concepts used in CG

Vectors

and Matrices

Basic

geometrical
concepts

u

au

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 2

LA Applications in AR

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 3

What we should know about Linear Algebra

3D

Coordinate geometry

Vectors
Dot

in 2 space and 3 space

product and cross product definitions and uses

Vector

and Matrix notation and algebra

Properties
Matrix

(matrix associativity but NOT matrix commutativity)

transpose and inverse definition, use and calculation

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 4

History of algebras.

Pamplona, V.F. 2006. Geometric Algebra brief introduction. 123, http://vitorpamplona.com/

wiki/Curriculum

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 5

Vector and Matrix notation: a non-geometric

example I

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 6

Vector and Matrix notation: a non-geometric

example II

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 7

Vector and Matrix notation: a non-geometric

example III

What

do I pay?

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 8

Vector and Matrix notation: a non-geometric

example IV

Using

Matrix notation:

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 9

Cartesian
coordinate space

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 10

Vectors & vector space I

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 11

Vectors & vector space II

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 12

Vector addition

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 13

Adding vectors
visually

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 14

Scalar
Multiplication I

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 15

Scalar
Multiplication II:
Linear
Dependence

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 16

Basis vectors
of the plane

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 17

Euclidean space

A

vector n in the n-dimensional Euclidean space

n

v0
v1
...
...
...
vn 1

with vi

, i 0, ... n

is denoted:

vector can also be represented as row vector

Most CG books and APIs use column-major form

Addition and multiplication by scalar:

u

u0
u1
...
...
...
un1

v0
v1
...
...
...
vn1

u0 v0
u1 v1
...
...
...
un1 vn1

av

au0
au1
...
...
...
aun 1

nmeans addition and multiplication yield vectors of the same

space

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 18

Coordinate Systems

Right

handed coordinate system

x
z
Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 19

Vector Arithmetic Summary

[
b = [b

a = ax

ay

by

[
a b = [a
a = [ a
sa = [sa

]
b]
az
z

]
b ]

a + b = a x + bx

a y + by

a z + bz

bx

a y by

az

x
x

ay
say

az
saz

Interactive Computer Graphics, Spring 2013, University of Crete

]
Lecture 2: Linear Algebra in CG Slide 20

Vector Magnitude

The

magnitude (length) of a vector is:

2
x

2
y

v = v +v +v

2
z

vector with length=1.0 is called a unit vector

We can also normalize a vector to make it a unit vector:

v
v
Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 21

Dot Product

a b = ai bi
a b = a x bx + a y by + a z bz
a b = a b cos

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 22

Dot Product

a b = ai bi
a b = a x bx + a y by + a z bz
a b = a b cos
T

ab = a b

a b = ax

Interactive Computer Graphics, Spring 2013, University of Crete

ay

bx

a z by
bz

Lecture 2: Linear Algebra in CG Slide 23

Example: Angle Between Vectors

How

do you find the angle between vectors a and b?

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 24

Example: Angle Between Vectors

a b = a b cos
a b

cos =

a
b

b
1

= cos

a
b

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 25

Dot Products with General Vectors

The

dot product is a scalar value that tells us something about the

relationship between two vectors

If

ab > 0 then < 90

If

ab < 0 then > 90

If

ab = 0 then = 90 (or one or more of the vectors is degenerate

(0,0,0))

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 26

Dot Products with One Unit Vector

If

|u|=1.0 then au is the length of the projection of a onto

u

au

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 27

Example: Distance to Plane

A

plane is described by a point p on the plane and a unit

normal n. Find the distance from point x to the plane

x

n

p


Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 28

Example: Distance to Plane

The

distance is the length of the projection of x-p onto n:

dist = (x p) n

x

n

x-p

p


Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 29

Dot Products with Unit Vectors

0 < ab <
ab = 0
1

ab = 1

b

a

-1 < ab < 0

ab = -1

Interactive Computer Graphics, Spring 2013, University of Crete

ab

a = b = 1.0
a b = cos( )
Lecture 2: Linear Algebra in CG Slide 30

Cross Product

i
a b = ax
bx

j
ay
by

k
az
bz

a b = a y bz a z by

a z bx a x bz

Interactive Computer Graphics, Spring 2013, University of Crete

a x by a y bx

Lecture 2: Linear Algebra in CG Slide 31

Properties of the Cross Product

a b is a vector perpendicular to both
a and b, in the direction defined by
the right hand rule

a b = a b sin
a b = area of parallelogram ab
a b = 0 if a and b are parallel
Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 32

Example: Normal of a Triangle

Find

the unit length normal of the triangle defined by 3D points

a, b, and c

a

Interactive Computer Graphics, Spring 2013, University of Crete

b

Lecture 2: Linear Algebra in CG Slide 33

Example: Normal of a Triangle

n = (b a ) (c a )

n
n=
n

c

c-a

a

b-a

Interactive Computer Graphics, Spring 2013, University of Crete

b

Lecture 2: Linear Algebra in CG Slide 34

Example: Area of a Triangle

Find

the area of the triangle defined by 3D points a, b, and c

a

Interactive Computer Graphics, Spring 2013, University of Crete

b

Lecture 2: Linear Algebra in CG Slide 35

Example: Area of a Triangle

1
area = (b a ) (c a )
2
c

c-a

a

b-a

Interactive Computer Graphics, Spring 2013, University of Crete

b

Lecture 2: Linear Algebra in CG Slide 36

Example: Alignment to Target

An

object is at position p with a unit length heading of h.

We want to rotate it so that the heading is facing some
target t. Find a unit axis a and an angle to rotate around.

h

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 37

Example: Alignment to Target

h ( t p)
a=
h ( t p)
a

t-p

$ h ( t p) '
1
)
= cos &&
)
h

p
(
)
%
(

t

h

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 38

Uses of the dot product

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 39

Rule for Dot product

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 40

Finding the length of a vector

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 41

Finding the
angle between
two vectors

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 42

More uses of
the Dot product

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 43

Summary: dot product and bases

Dot

product

u
v

Orthonormal

cos

basis

0, i j,
1, i j.

u
i uj
if

the basis vectors are mutually perpendicular, but not of unit length:
orthogonal

Orthonormal

Project
w

does not have to be simple vectors e.g. spherical harmonics

a vector orthogonally onto another vector:

uv
v

uv
vv

tv

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 44

 Vector

cross product I

(cross) product

If

w
w

|uv|=|u||v|sin
u and w

u is parallel to v, cross product is 0

u,v,w

form a right-handed system

Properties

u v v u
au
bv w a u w
b v w
u v w v w u
u v w u w v u v w

Interactive Computer Graphics, Spring 2013, University of Crete

anti commutativity
linearity
scalar triple product
vector triple product

Lecture 2: Linear Algebra in CG Slide 45

Cross product
II

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 46

Linear Algebra: vector rules

(u+v)
u+v

+ w = u + (v + w) (associativity)

= v+u (commutativity)

+ v = v (zero identity, with zero vector 0 = (0,0,.0) )

+ (-v) = 0 (additive inverse)

Multiplication
(ab)u

by a scalar rules:

= a(bu)

(a+b)u

= au + bu (distributive law)

a(u+v)

= au + av (distributive law)

1u=u

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 47

Problem: Translating objects

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 48

What is a Matrix?

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 49

Translation

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 50

Scaling

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 51

Rotation

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 52

Sets of Linear
Equations and
Matrices

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 53

Matrix-Vector Multiplication I

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 54

Matrix-Vector Multiplication II

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 55

Algebraic
Properties of
Matrices

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 56

Matrix
multiplication
analyzed

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 57

Matrix
Inverse

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 58

What does
an Inverse
do?

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 59

Matrices (1/2)

A

tool for manipulating vectors 

and points

Identity

m00
m01
m10
m11
...
...
mp 1,0 mp 1,1

... m0,q 1
... m1,q 1
...
...
... mp 1,q 1

mij

matrix

Matrix-matrix

addition

Scalar-matrix

multiplication

Transpose:

1
0 ... 0
0
1 ... 0
... ... ... ...
0
0 ... 0
0
0 ... 1

M=[mij], MT=[mji]

M

Interactive Computer Graphics, Spring 2013, University of Crete

mij

nij mij

nij

Lecture 2: Linear Algebra in CG Slide 60

Matrices (2/2)

Matrix-Matrix
M

of size pxq and N qxr , T becomes pxr

m00
m10
...

T MN

mp

1,0

m01
m11
...
mp

1,1

Determinant
M

Inverse

n00
n10
...
np

n01
n11
...
np

1,0

... n0,r 1
... n1,r 1
...
...
... np 1,r 1

1,1

q 1
i0

q 1
i0

m0,i ni,0
...
...

mp

1,i

...
...
...

ni,0 ...

q 1
i0

q 1
i0

m0,i ni,r
...
...

mp

1,i

ni,r

of matrix M

m00 m11

m01 m10

of matrix M

Co-factor d

... m0,q 1
... m1,q 1
...
...
... mp 1,q 1

m00 m01
m10 m11

Adjoint

multiplication

dM02

m00 m01 m02

m10 m11 m12
m20 m21 m22

M
m10 m11
m20 m21

adj M

d00
d01
d02

d10
d11
d12

d20
d21
d22

of Matrix (use the formula below-left or Gauss-Jordan elimination below-right)

1
M

adj M
M

MI

MT

1 T

MN

Interactive Computer Graphics, Spring 2013, University of Crete

1
1

Lecture 2: Linear Algebra in CG Slide 61

Orthogonal Matrices

A

square matrix M, with only real elements and

T

MT M I
M 1

MM
M

MT

an orthogonal matrix is not the same as an orthogonal vector set

(mutually perpendicular vectors)

An
A

orthogonal matrix must consist of normalized vectors

rotation is an orthogonal matrix

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 62

Matrix
Notation

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 63

But, there is a problem

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 64

Solution to
notational
problem

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 65

Matrix Notation and Composition

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 66

Basic Trigonometry for CG I

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 67

Basic Trigonometry for CG II

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 68

Geometry: lines

r(t)

= o +td

is a point on the line

is direction of the line

is a parameter to generate
different points r on the line

ax+by+c=0

For

p to lie on lie L it must 

fulfill this equation

If

a,b are combined in vector

n=(nx,ny)=(a,b)

n.p

+ c =0

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 69

Geometry: Planes

Explicit

form: P(u,v)=o +us+vt

o is a point lying on the plane

s, t are direction vectors that span the plane (i.e. noncollinear)

u,v parameters to generate different points on the plane

Plane normal is sxt

Implicit

form: n.p + d=0

n is plane normal

P is a point on the plane

d is a constant that determines part of the position of the plane

f(p)=n.p + d

If f(p)=0, then p is on the plane

If f(p)>0, then p lies on the same side of the plane as q+n

If f(p)<0, then p lies on the same side of the plane as q-n

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 70

Geometry: Convex Hull

The

smallest set that

satisfies the following
condition:

The

straight line between

any two points in the set
is totally included in the set as well

Rubber
Used

band example

for bounding volume algorithms

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 71

References

Dam,

A. 2003. Linear Algebra. Introduction to Computer

Graphics, CS Brown University, 153.

Akenine-Mller,

T., Haines, E., and Hoffman, N. 2008.

Appendix A. Some linear Algebra. Real-Time Rendering, 123.

Akenine-Mller,

T., Haines, E., and Hoffman, N. 2008.

Appendix B. Trigonometry. Real-Time Rendering, 17.

Rotenberg,

S., 2005, Vectors & Matrices, Computer Graphics,

UCSD

Interactive Computer Graphics, Spring 2013, University of Crete

Lecture 2: Linear Algebra in CG Slide 72

Thank you!

Interactive Computer Graphics, Spring 2013, University of Crete