Geometric transformation

2D Transformation

Basic transformations Matrix Representations and Homogeneous Coordinates Composite Transformations Other transformations : Reflection & Shear Transformations Between Coordinate systems

Why Transformations?
In graphics, once we have an object described, transformations are used to move that object, scale it and rotate it

2D Transformation

Given a 2D object, transformation is to change the objects

Position (translation) Size (scaling) Orientation (rotation) Shapes (shear)

Perform Transformation on any Object:

[transformed Object Matrix]= [original Object Matrix]*[Transformation Matrix] OR [X*]=[X][T]

Point representation
Two way to represent point in 2D 1. column vector (a 2x1 matrix) to represent a 2D point x y 2. Row Vector(a 1x2 matrix)

A general form of linear transformation can be written as: x = ax + by + c y = dx + ey + f OR

X Y 1 a d 0 b e 0 c f 1 x y 1

Translation

Re-position a point along a straight line Given a point (x,y), and the translation distance (tx,ty)
The new point: (x, y) x = x + tx y = y + ty
(x,y)
ty

(x,y)

tx

OR

P = P + T

where

P =

x y

p=

x y

T=

tx ty

3x3 2D Translation Matrix

x y = x y + tx ty

Use 3 x 1 vector

x y 1

1 0 0

0 1 0

tx ty 1

x y 1

Note that now it becomes a matrix-vector multiplication

Translation

How to translate an object with multiple vertices?

Translate individual vertices

Transformation of straight line

Let us consider the coordinates of end point of straight line AB To be A=[0 2] and B=[3 2]. The elements of transformation matrix [T] may be assumed. So the effect can be shown as [L][T]=L*

Midpoint transformation

Transformation of parallel lines

Transformation of intersecting lines

2D Rotation

Default rotation center: Origin (0,0)

> 0 (positive ): Rotate counter clockwise

< 0(negative) : Rotate clockwise

Rotation
(x,y) -> Rotate about the origin by q (x, y) How to compute (x, y) ? x = r cos (f) y = r sin (f) x = r cos (f + q) y = r sin (f + q)
q (x,y)

r
f

(x,y)

Rotation
x = r cos (f) y = r sin (f) x = r cos (f + q) y = r sin (f + q) x = r cos (f + q) = r cos(f) cos(q) r sin(f) sin(q) = x cos(q) y sin(q) y = r sin (f + q) = r sin(f) cos(q) + r cos(f)sin(q) = y cos(q) + x sin(q)
q (x,y)

r
f

(x,y)

Rotation
x = x cos(q) y sin(q) y = y cos(q) + x sin(q) Matrix form? x y 3 x 3?
= q (x,y)

r
f

(x,y)

cos(q) sin(q)

-sin(q) cos(q)

x y

3x3 2D Rotation Matrix

x y
=

cos(q) sin(q)

-sin(q) cos(q)

x y
q

(x,y) (x,y) f

x y 1

cos(q) sin(q) 0

-sin(q) cos(q) 0

0 0 1

x y 1

Rotation

How to rotate an object with multiple vertices?

Rotate individual Vertices

Reflection

The reflection about x axis i.e. y=0

-

The reflection about y axis i.e. x=0 For reflection about the line y=x For reflection about the line y=-x

2D Scaling
Scale: Alter the size of an object by a scaling factor (Sx, Sy), i.e. x = x . Sx y = y . Sy x y = Sx 0 0 Sy x y
(4,4) (2,2) (1,1) Sx = 2, Sy = 2 (2,2)

2D Scaling
(4,4) (2,2) (1,1) Sx = 2, Sy = 2 (2,2)

Not only the object size is changed, it also moved!! Usually this is an undesirable effect We will discuss later (soon) how to fix it

3x3 2D Scaling Matrix

x y = Sx 0 0 Sy x y

x y 1

Sx 0 0

0 Sy 0

0 0 1

x * y 1

Put it all together

Translation: x = x y y Rotation:

tx ty

x = cos(q) -sin(q) * x y sin(q) cos(q) y x y

=

Scaling:

Sx 0

0 Sy

x y

Or, 3x3 Matrix representations

Translation:

x y 1 x y 1 x y 1

1 0 0 cos(q) sin(q) 0 Sx 0 0 0 Sy 0

0 1 0

tx ty 1

x y 1

Rotation:

-sin(q) cos(q) 0 0 0 1

0 x 0 * y 1 1 x * y 1

Scaling:

Why use 3x3 matrices?

Why use 3x3 matrices?

So that we can perform all transformations using matrix/vector multiplications This allows us to pre-multiply all the matrices together The point (x,y) needs to be represented as (x,y,1) -> this is called Homogeneous coordinates!

Rotation Revisit

The standard rotation matrix is used to rotate about the origin (0,0)
cos(q) sin(q) 0 -sin(q) cos(q) 0 0 0 1

What if I want to rotate about an arbitrary center?

Arbitrary Rotation Center

To rotate about an arbitrary point P (px,py) by q:

Translate the object so that P will coincide with the origin: T(-xr, -yr) Rotate the object: R(q) Translate the object back to its original position: T(xr,yr)

(px,py)

Arbitrary Rotation Center

Translate the object so that P will coincide with the origin: T(-px, -py) Rotate the object: R(q) Translate the object back: T(px,py)
T(px,py) R(q) T(-px, -py)

Put in matrix form:

x y = 1 1 0 px 0 1 py 00 1

*P
x y 1

cos(q) -sin(q) 0 sin(q) cos(q) 0 0 0 1

1 0 -px 0 1 -py 0 0 1

Reflection about an arbitrary line

Translate the line as well as the object so that line passes through the origin. Rotate the line and object about the origin Reflect the object through the coordinate axis Apply the inverse rotation about the origin to shift the line at translated position Apply inverse translation , move line to its original position.

Scaling Revisit

The standard scaling matrix will only anchor at (0,0)

Sx 0 0 0 Sy 0 0 0 1

What if I want to scale about an arbitrary pivot point?

Arbitrary Scaling Pivot

To scale about an arbitrary pivot point P (px,py):

Translate the object so that P will coincide with the origin: T(-px, -py) Rotate the object: S(sx, sy) Translate the object back: T(px,py)

(px,py)

Shearing

A transformation that distorts the shape of an object such that transformed shape appears as if object were composed of internal layers that had been caused to slide over each other is called shearing.

1. 2.

Two cases: x-direction shear Y-direction shear

1. Shearing in x

Y coordinates are unaffected, but x cordinates are translated linearly with y That is: x 1 shx 0 x

y = y x = x + shx * h

y = 1

0 1 0 * y 0 0 1 1

Shearing in y
x y = 1 1 0 0 x shy 1 0 * y 0 0 1 1

X coordinates are unaffected, but y cordinates are translated linearly with y

Interesting Facts:

A 2D rotation is three shears Shearing will not change the area of the object Any 2D shearing can be done by a rotation, followed by a scaling, and followed by a rotation

Affine Transformation

Translation, Scaling, Rotation, Shearing are all affine transformation Affine transformation transformed point P (x,y) is a linear combination of the original point P (x,y), i.e.
x m11 m12 m13 y = m21 m22 m23 1 0 0 1 x y 1

Any 2D affine transformation can be decomposed into a rotation, followed by a scaling, followed by a shearing, and followed by a translation.
Affine matrix = translation x shearing x scaling x rotation

Composing Transformation

Composing Transformation the process of applying several transformation in succession to form one overall transformation If we apply transform a point P using M1 matrix first, and then transform using M2, and then M3, then we have: (M3 x (M2 x (M1 x P ))) = M3 x M2 x M1 x P
(pre-multiply)

Composing Transformation

Matrix multiplication is associative

M3 x M2 x M1 = (M3 x M2) x M1 = M3 x (M2 x M1)

Transformation products may not be commutative xA Some cases where A x B = B x A

A translation scaling rotation uniform scaling (sx = sy) B translation scaling rotation rotation

A x B != B

Transformation order matters!

Example: rotation and translation are not commutative

Translate (5,0) and then Rotate 60 degree OR Rotate 60 degree and then translate (5,0)??

Rotate and then translate !!