Fractal Geometry Methods

Sunu Fathima T.H

1
Fractal Geometry Methods
❏ Previous object representations:
❏ Object shapes were described with equations.
❏ Have smooth surfaces and regular shapes.
❏ Natural objects, such as mountains and clouds
❏ Have irregular or fragmented features
❏ Euclidean methods do not provide realistic representations for such
objects.
❏ Natural objects can be realistically described with fractal-geometry
methods.
❏ Procedures rather than equations are used to model objects

2
Fractal Geometry Methods
❏ A fractal object has two basic characteristics:
❏ Infinite detail at every point
❏ A certain self-similarity between the object parts and the overall features
of the object
❏ We describe a fractal object with a procedure that specifies a repeated
operation for producing the detail in the object subparts.
❏ Natural objects are represented with procedures that theoretically repeat an
infinite number of times.

3
Fractal Geometry Methods
❏ If we zoom in on a fractal object, we continue to see more and more details
in the magnifications without an eventual smoothing of the object
appearance.
❏ A mountain outlined against the sky continues to have the same jagged shape
as we view it from a closer and closer position.
❏ As we near the mountain, the smaller detail in the individual ledges and
boulders becomes apparent.

4
Fractal Geometry Methods

5
Fractal Geometry Methods
❏ Can characterize the amount of variation in object detail with a number
called the fractal dimension.
❏ Sometimes referred to as the fractional dimension, which is the basis for
the name “fractal.”
❏ Useful for modeling a very wide variety of natural phenomena.
❏ Used to model terrain, clouds, water, trees and other plant life, feathers, fur,
and various surface textures.
❏ Sometimes just to make pretty patterns.

6
Fractal Generation Procedures
❏ A fractal object is generated by repeatedly applying a specified
transformation function to points within a region of space.
❏ If P0 = (x0, y0, z0) is a selected initial position, each iteration of a
transformation function F generates successive levels of detail with the
calculations:
P1 = F (P0), P2 = F (P1), P3 = F (P2), ···

7
Classification of Fractals
❏ Self-similar fractals have parts that are scaled-down versions of the entire object.
❏ Starting with an initial shape, construct the object subparts by apply a scaling
parameter s to the overall shape.
❏ Use the same scaling factor s for all subparts, or we can use different scaling
factors for different scaled-down parts of the object.
❏ Apply random variations to the scaled-down subparts, the fractal is said to be
statistically self-similar.
❏ The parts then have the same statistical properties.
❏ Statistically self-similar fractals are commonly used to model trees, shrubs,
and other vegetation.

8
Classification of Fractals
❏ Self-affine fractals have parts that are formed with different scaling
parameters, sx, sy, and sz, in different coordinate directions.
❏ Include random variations to obtain statistically self-affine fractals.
❏ Terrain, water, and clouds are typically modeled with statistically
self-affine fractal construction methods.
❏ Invariant fractal sets are formed with nonlinear transformations.
❏ Includes:
❏ self-squaring fractals: such as the Mandelbrot set (formed with
squaring functions in complex space)
❏ self-inverse fractals: constructed with inversion procedures

9
Fractal Dimension
❏ The amount of variation in the structure of a fractal object can be described
with a number D, called the fractal dimension
❏ A measure of the roughness, or fragmentation, of the object.
❏ More jagged-looking objects have larger fractal dimensions.

10
Geometric Construction of Deterministic Self-Similar Fractals
❏ To construct a deterministic (nonrandom) self-similar fractal geometrically:
❏ Start with a given geometric shape, called the initiator.
❏ Subparts of the initiator are then replaced with a pattern, called the
generator.

11
Geometric Construction of Deterministic Self-Similar Fractals
❏ Example: snowflake pattern, or Koch curve
❏ Each straight-line segment in the initiator is replaced with the generator
pattern, consisting of four equal-length line segments.
❏ Then the generator is scaled and applied to the line segments of the modified
initiator.
❏ This process is repeated for some number of steps.
❏ The scaling factor at each step is 1/3

12
Geometric Construction of Deterministic Self-Similar Fractals

13
Geometric Construction of Deterministic Self-Similar Fractals

14
Geometric Construction of Deterministic Self-Similar Fractals
❏ Displays of trees and other plants can be constructed with self-similar
geometric-construction methods.

15
Geometric Construction of Statistically Self-Similar Fractals
❏ Introduce variability into the geometric construction of a self-similar fractal.
❏ Randomly select a generator at each step from a menu of patterns.
❏ Compute coordinate displacements with small random variations.
❏ Example: Use a probability distribution function to compute variable
midpoint displacements at each step in the creation of a random snowflake
pattern.

16
Geometric Construction of Statistically Self-Similar Fractals

17
Geometric Construction of Statistically Self-Similar Fractals
❏ Random scaling parameters and branching directions are used in this display
to model the vein patterns in a leaf.

18
Geometric Construction of Statistically Self-Similar Fractals
❏ Once a fractal object has been created, we can model a scene using several
transformed instances of the object.
❏ Instancing with random rotations of a fractal tree.
❏ A fractal forest can be displayed using various random transformations.

19
Affine Fractal-Construction Methods
❏ Obtain highly realistic representations for terrain and other natural objects.
❏ Model object features as fractional Brownian motion
❏ Extension of standard Brownian motion, a form of “random walk,” which
describes the erratic, zigzag movement of particles in a gas or fluid.

20
Affine Fractal-Construction Methods
❏ Starting from a given position, we generate a straight-line segment in a
random direction and with a random length.
❏ Another random line is then constructed from the endpoint of this first line.
❏ Process is repeated for a designated number of line segments.
❏ Fractional Brownian motion is obtained by adding an additional parameter to
the statistical distribution describing Brownian motion.
❏ This additional parameter sets the fractal dimension for the “motion” path.

21
Self-Squaring Fractals
❏ Apply a transformation function repeatedly to points in complex space.
❏ In two dimensions, a complex number can be represented as z = x + iy, where
x and y are real numbers and i2 = −1.
❏ In three-dimensional and four-dimensional space, points are represented with
quaternions.
❏ A complex squaring function f (z) is one that involves the calculation of z2,
and we can use some self-squaring functions to generate fractal shapes.

22
Self-Squaring Fractals
❏ Depending on the initial position selected for the iteration, repeated
application of a self-squaring function will produce one of three possible
outcomes:
❏ The transformed position can diverge to infinity.
❏ The transformed position can converge to a finite limit point, called an
attractor.
❏ The transformed position remains on the boundary of a region.

23
Self-Squaring Fractals

24
Self-Squaring Fractals
❏ Non-fractal squaring operation f (z) = z2 in the complex plane transforms positions
according to their relation to the unit circle.
❏ Any point z whose magnitude |z| is greater than 1 is transformed through a
sequence of positions that tend to infinity.
❏ A point with |z| < 1 is transformed toward the coordinate origin.
❏ Points that are originally on the circle, |z| = 1, remain on the circle.
❏ Although the z2 transformation does not produce a fractal, some complex squaring
operations generate a fractal curve as the boundary between those positions that
move toward infinity and those that tend toward a finite limit.
❏ A closed fractal boundary generated with a squaring operation is called a Julia set

25
Self-Inverse Fractals
❏ Various geometric inversion transformations can be used to create fractal
shapes.
❏ Start with an initial set of points.
❏ Repeatedly apply nonlinear inversion operations to transform the initial
points into a fractal.
❏ For eg: we consider a two-dimensional inversion transformation with respect
to a circle of radius r and center position Pc = (xc , yc ).
❏ A point P outside the circle is inverted to a position P inside the circle with
the transformation
❏ Both P and P’ lie on a straight line passing through the circle center Pc

26
Self-Inverse Fractals

27
Self-Inverse Fractals
❏ We can apply this transformation to various objects, such as straight lines,
circles, or ellipses.
❏ A straight line that passes through the circle center is invariant under this
inversion transformation; it maps into itself.
❏ A straight line that does not pass through the circle center inverts into a circle
whose circumference contains the center point Pc
❏ Any circle that passes through the center of the reference circle is inverted
into a straight line that does not pass through the circle center.
❏ If the circle does not intersect the center of the reference circle, it inverts into
another circle

28
Self-Inverse Fractals
❏ Another invariant inversion is the transformation of a circle that is orthogonal to the
reference circle.
❏ That is, the tangents of the two circles are perpendicular at the intersection points.
❏ We can create various fractal shapes with this inversion transformation by starting
with a set of circles and repeatedly applying the transformation using different
reference circles.
❏ Similarly, we can apply circle inversion with respect to a set of straight lines.
❏ Comparable inversion methods can be developed for other two-dimensional shapes.
❏ In addition, we can generalize the procedure to spheres, planes, or other
three-dimensional objects.

29
Self-Inverse Fractals

30
Thank You

31