Transformations

MIT EECS 6.837, Durand and Cutler

Outline
• Assignment 0 Recap
• Intro to Transformations
• Classes of Transformations
• Representing Transformations
• Combining Transformations
• Change of Orthonormal Basis

MIT EECS 6.837, Durand and Cutler

Cool Results from Assignment 0

emmav alexg psoto

koi scyudits
MIT EECS 6.837, Durand and Cutler
Notes on Assignments
• Make sure you turn in a linux or windows
executable (so we can test your program)
• Don't use athena dialup
• Collaboration Policy
– Share ideas, not code
• Ask questions during office hours, or email
6.837-staff@graphics.csail.mit.edu
• Tell us how much time you spent on each
assignment

MIT EECS 6.837, Durand and Cutler

Quick Review of Last Week
• Ray representation
• Generating rays from eyepoint / camera
– orthographic camera
– perspective camera
• Find intersection point & surface normal
• Primitives:
– spheres, planes, polygons, triangles, boxes

MIT EECS 6.837, Durand and Cutler

Outline
• Assignment 0 Recap
• Intro to Transformations
• Classes of Transformations
• Representing Transformations
• Combining Transformations
• Change of Orthonormal Basis

MIT EECS 6.837, Durand and Cutler

What is a Transformation?
• Maps points (x, y) in one coordinate system to
points (x', y') in another coordinate system
x' = ax + by + c
y' = dx + ey + f

• For example, IFS:

MIT EECS 6.837, Durand and Cutler

Common Coordinate Systems
• Object space
– local to each object
• World space
– common to all objects
• Eye space / Camera space
– derived from view frustum
• Screen space
– indexed according
to hardware attributes

MIT EECS 6.837, Durand and Cutler

Simple Transformations

• Can be combined
• Are these operations invertible?
Yes, except scale = 0

MIT EECS 6.837, Durand and Cutler

Transformations are used:
• Position objects in a scene (modeling)
• Change the shape of objects
• Create multiple copies of objects
• Projection for virtual cameras
• Animations

MIT EECS 6.837, Durand and Cutler

Outline
• Assignment 0 Recap
• Intro to Transformations
• Classes of Transformations
• Rigid Body / Euclidean Transforms
• Similitudes / Similarity Transforms
• Linear
• Affine
• Projective
• Representing Transformations
• Combining Transformations
• Change of Orthonormal Basis
MIT EECS 6.837, Durand and Cutler
Rigid-Body / Euclidean Transforms

• Preserves distances
• Preserves angles

Rigid / Euclidean

Identity
Translation
Rotation

MIT EECS 6.837, Durand and Cutler

Similitudes / Similarity Transforms

• Preserves angles

Similitudes
Rigid / Euclidean

Identity
Translation Isotropic Scaling
Rotation

MIT EECS 6.837, Durand and Cutler

Linear Transformations

Similitudes
Linear
Rigid / Euclidean
Scaling
Identity
Translation Isotropic Scaling Reflection
Rotation
Shear

MIT EECS 6.837, Durand and Cutler

Linear Transformations
• L(p + q) = L(p) + L(q)
• L(ap) = a L(p)

Similitudes
Linear
Rigid / Euclidean
Scaling
Identity
Translation Isotropic Scaling Reflection
Rotation
Shear

MIT EECS 6.837, Durand and Cutler

Affine Transformations
• preserves
parallel lines
Affine

Similitudes
Linear
Rigid / Euclidean
Scaling
Identity
Translation Isotropic Scaling Reflection
Rotation
Shear

MIT EECS 6.837, Durand and Cutler

Projective Transformations
• preserves lines
Projective
Affine

Similitudes
Linear
Rigid / Euclidean
Scaling
Identity
Translation Isotropic Scaling Reflection
Rotation
Shear

Perspective
MIT EECS 6.837, Durand and Cutler
Perspective Projection

MIT EECS 6.837, Durand and Cutler

Outline
• Assignment 0 Recap
• Intro to Transformations
• Classes of Transformations
• Representing Transformations
• Combining Transformations
• Change of Orthonormal Basis

MIT EECS 6.837, Durand and Cutler

How are Transforms Represented?

x' = ax + by + c
y' = dx + ey + f

x' a b x c
= +
y' d e y f

p' = Mp + t
MIT EECS 6.837, Durand and Cutler
Homogeneous Coordinates
• Add an extra dimension
• in 2D, we use 3 x 3 matrices
• In 3D, we use 4 x 4 matrices
• Each point has an extra value, w
x' a b c d x
y' e f g h y
=
z' i j k l z
w' m n o p w

p' = Mp
MIT EECS 6.837, Durand and Cutler
Homogeneous Coordinates
• Most of the time w = 1, and we can ignore it

x' a b c d x
y' e f g h y
=
z' i j k l z
1 0 0 0 1 1

• If we multiply a homogeneous coordinate

by an affine matrix, w is unchanged

MIT EECS 6.837, Durand and Cutler

Homogeneous Visualization
• Divide by w to normalize (homogenize)
• W = 0? Point at infinity (direction)

(0, 0, 1) = (0, 0, 2) = … w = 1
(7, 1, 1) = (14, 2, 2) = …
(4, 5, 1) = (8, 10, 2) = … w=2
MIT EECS 6.837, Durand and Cutler
Translate (tx, ty, tz) Translate(c,0,0)
y

• Why bother with the

extra dimension? p p'
Because now translations
x
can be encoded in the matrix! c

x' 1 0 0 tx x
y' 0 1 0 ty y
=
z' 0 0 1 tz z
10 0 0 0 1 1

MIT EECS 6.837, Durand and Cutler

Scale (sx, sy, sz) Scale(s,s,s)
y p'

• Isotropic (uniform)
p
scaling: sx = sy = sz q'
q
x

x' sx 0 0 0 x
y' 0 sy 0 0 y
=
z' 0 0 sz 0 z
1 0 0 0 1 1

MIT EECS 6.837, Durand and Cutler

Rotation ZRotate(θ)
y
p'
• About z axis
θ p

x
z

x' cos θ -sin θ 0 0 x

y' sin θ cos θ 0 0 y
=
z' 0 0 1 0 z
1 0 0 0 1 1

MIT EECS 6.837, Durand and Cutler

Rotation
x' 1 0 0 0 x
• About y' 0 cos θ -sin θ 0 y
x axis: =
z' 0 sin θ cos θ 0 z
1 0 0 0 1 1

x' cos θ 0 sin θ 0 x

• About y' 0 1 0 0 y
=
y axis: z' -sin θ 1 cos θ 0 z
1 0 0 0 1 1
MIT EECS 6.837, Durand and Cutler
 Rotation y
Rotate(k, θ)

• About (kx, ky, kz), a unit θ

vector on an arbitrary axis k
(Rodrigues Formula) x
z

x' kxkx(1-c)+c kzkx(1-c)-kzs kxkz(1-c)+kys 0 x

y' kykx(1-c)+kzs kzkx(1-c)+c kykz(1-c)-kxs 0 y
=
z' kzkx(1-c)-kys kzkx(1-c)-kxs kzkz(1-c)+c 0 z
1 0 0 0 1 1

where c = cos θ & s = sin θ

MIT EECS 6.837, Durand and Cutler
Outline
• Assignment 0 Recap
• Intro to Transformations
• Classes of Transformations
• Representing Transformations
• Combining Transformations
• Change of Orthonormal Basis

MIT EECS 6.837, Durand and Cutler

How are transforms combined?
Scale then Translate
(5,3)
Scale(2,2) (2,2) Translate(3,1)
(1,1) (3,1)
(0,0) (0,0)

Use matrix multiplication: p' = T ( S p ) = TS p

1 0 3 2 0 0 2 0 3
TS = 0 1 1 0 2 0 = 0 2 1
0 0 1 0 0 1 0 0 1

Caution: matrix multiplication is NOT commutative!

MIT EECS 6.837, Durand and Cutler
Non-commutative Composition
Scale then Translate: p' = T ( S p ) = TS p
(5,3)
Scale(2,2) (2,2) Translate(3,1)
(1,1) (3,1)
(0,0) (0,0)

Translate then Scale: p' = S ( T p ) = ST p

(8,4)
Translate(3,1) (4,2) Scale(2,2)
(1,1) (6,2)
(3,1)
(0,0)

MIT EECS 6.837, Durand and Cutler

Non-commutative Composition
Scale then Translate: p' = T ( S p ) = TS p
1 0 3 2 0 0 2 0 3
TS = 0 1 1 0 2 0 = 0 2 1
0 0 1 0 0 1 0 0 1

Translate then Scale: p' = S ( T p ) = ST p

2 0 0 1 0 3 2 0 6
ST = 0 2 0 0 1 1 = 0 2 2
0 0 1 0 0 1 0 0 1

MIT EECS 6.837, Durand and Cutler

Outline
• Assignment 0 Recap
• Intro to Transformations
• Classes of Transformations
• Representing Transformations
• Combining Transformations
• Change of Orthonormal Basis

MIT EECS 6.837, Durand and Cutler

Review of Dot Product

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis
• Given: y
v x p
coordinate frames u
v
u
xyz and uvn y
point p = (x,y,z) x

y
• Find: v
p u
p = (u,v,n)
n x
z

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis
y y
v y.u v
y.v
u u
x.v
x n x

x.u
z

x = (x . u) u + (x . v) v + (x . n) n
y = (y . u) u + (y . v) v + (y . n) n
z = (z . u) u + (z . v) v + (z . n) n

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis
x = (x . u) u + (x . v) v + (x . n) n
y = (y . u) u + (y . v) v + (y . n) n
z = (z . u) u + (z . v) v + (z . n) n

Substitute into equation for p:

p = (x,y,z) = x x + y y + z z
p = x [ (x . u) u + (x . v) v + (x . n) n ] +
y [ (y . u) u + (y . v) v + (y . n) n ] +
z [ (z . u) u + (z . v) v + (z . n) n ]

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis

p = x [ (x . u) u + (x . v) v + (x . n) n ] +
y [ (y . u) u + (y . v) v + (y . n) n ] +
z [ (z . u) u + (z . v) v + (z . n) n ]

Rewrite:
p = [ x (x . u) + y (y . u) + z (z . u) ] u +
[ x (x . v) + y (y . v) + z (z . v) ] v +
[ x (x . n) + y (y . n) + z (z . n) ] n

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis
p = [ x (x . u) + y (y . u) + z (z . u) ] u +
[ x (x . v) + y (y . v) + z (z . v) ] v +
[ x (x . n) + y (y . n) + z (z . n) ] n
p = (u,v,n) = u u + v v + n n
Expressed in uvn basis:
u = x (x . u) + y (y . u) + z (z . u)
v = x (x . v) + y (y . v) + z (z . v)
n = x (x . n) + y (y . n) + z (z . n)

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis

u = x (x . u) + y (y . u) + z (z . u)
v = x (x . v) + y (y . v) + z (z . v)
n = x (x . n) + y (y . n) + z (z . n)

In matrix form:
where:
u ux uy uz x ux = x . u
v = vx vy vz y uy = y . u
n nx ny nz z etc.

MIT EECS 6.837, Durand and Cutler

Change of Orthonormal Basis
u ux uy uz x x
v = vx vy vz y = M y
n nx ny nz z z

What's M-1, the inverse?

x xu x v xn u u x = x . u = u . x = xu
y = yu y v yn v
z zu zv zn n
M = M
-1 T

MIT EECS 6.837, Durand and Cutler

Next Time:

Adding Transformations
to the Ray Caster
(Assignment 2)

MIT EECS 6.837, Durand and Cutler