September 26, 2022 8:5 0218-348X

Fractals, (15 pages)

© World Scientific Publishing Company

Hybrid Algorithm for the Classification of

Fractal Designs and Images

Zhenhua Yu1 , Ayesha Sohail2 , Maryam Jamil2 ,

O A Beg3 , João Manuel R.S. Tavares4a
1
Institute of Systems Security and Control, College of Computer Science and Technology, Xi’an University of Science and Technology, Xi’an
710054, China,
2
Department of Mathematics, Comsats University Islamabad, Lahore Campus 54000, Pakistan
3
Department of Mechanical/Aeronautical Engineering, SEE, University of Salford, Manchester, UK
4
Instituto de Ciência e Inovação em Engenharia Mecânica e Engenharia Industrial, Departamento de Engenharia Mecânica, Faculdade de
Engenharia, Universidade do Porto, Porto, Portugal.

Abstract
The Fractal patterns are recursive patterns and are self similar in nature. The
fractal geometry provides better understanding of natural patterns as com-
pared to the euclidean geometry. The fractal designs have been used exten-
sively in the fields of applied sciences due to the systematic methods used for
their generation. These methods provide benchmarks to analyze the rough-
ness, narrow/ broad vision of the objects. In the fields of architecture and
design, computational methods for fractal generation can prove to be more
reliable tools. The fractal patterns can be simulated and the architecture can
be modeled, with several options, before implementing it practically. During
this research, this strategy is opted to design novel fractal tile designs. Several
designs are selected from a series of simulations, based on the final visionary

a [email protected]

output-1
September 26, 2022 8:5 0218-348X

evaluation, according to the requirement of the walls of different zones in mod-

ern buildings. Four fractal patterns are simulated with several orientations and
final designs are documented with corresponding geometrical evaluation.

Keywords: Fractals; Tile design; Geometry processing; computational framework; Simulations.

 September 26, 2022 8:5 0218-348X

1. Introduction

The mathematical graphic designs and the resulting art has improved the decorating pat-
terns in the fields of architecture and building designs. Fractal patterns are used extensively
to design the tiles and walls. Art and mathematics have a close connection. Artists can use
mathematical models to create art. The resulting aesthetic factors play an important role in
the study of mathematics; thus, in both ways, mathematics is regarded as an art [1, 2, 3, 4].
The popularity of mathematical art graphics has helped to enhance the vitality of the tile
patterns, in order to decorate the floors, roofs and walls. As a result, the novel designs for
the tiles printing have attracted attention of the researchers and the improved variants can
help to boost the impression of the walls and thus the buildings. Architecture designs are
themselves symbols of the success of mathematics, such as the Fibonacci number and the
golden ratios were used by the Greeks in architecture and were prevalent in Renaissance
architecture. The applied mathematical programming is thus helpful in almost all domains
of science and technology [5, 6]. With the aid of mathematical models, programming and
simulations, the designs can be improved in novel ways.
The advancement of modern computer technology helped the tile pattern designers to
provide new resources for the tile patterns, decorative tile-work, and tile art.
There are numerous types of tile patterns that are made up of a large number of tiny
irregularly positioned, neutral components. Such components are usually made up of
glass but sometimes of ceramic or stone [7]. These patterns contribute to an architecture’s
finishing, worth, beauty and elegance. In history, Fractal patterns have been used to design
buildings [8, 9]. With the passage of time, this field of research evolved by the virtue of
modern programming tools [10, 11]. Fractal geometries in the history of art have served as
a language that has proven its worth through its applications. For this reason, during this
research a series of Fractal patterns is simulated using novel programming tools.
The history of fractal generation and its mathematical programming dates back since
the pioneering work of Mandelbrot, the famous American French mathematician. During
his research work on fractal geometry, he identified it as a subject that uses regular or
irregular geometry as the research object, in which nature is viewed as a fine structure with
infinite nesting levels and some self-similar properties at different scales [12]. Therefore,
the naturally occurring objects, such as trees, coastlines, and clouds have fractal properties,
September 26, 2022 8:5 0218-348X

and much of the prevailing interest in the subject stems from attempts to recreate such
natural phenomena using visual effects. The availability of modern computer graphics
techniques have provided new insight into the design of such patterns. Users can generate
fractals mathematically that can be reproduced at any magnification or reduction, and the
representation of each part looks exactly like the original, or at least has similar patterns [13].
Fractal tiling with self-similarity can generate more dynamic patterns for the buildings
and have a strong visual appearance. However, the fractal tilings that have been discovered
so far are very limited in number because creating such fractal tilings requires knowledge
of resulting percentages of colors used, image magnifications, geometrical aspects, special
talent and skill. This paper presents a general method for generating tile designs based on
the idea of hierarchically subdividing adjacent tiles [14].
Design patterns and tile designs are both the outcome of a systematic repetition of
a unique geometrical form in modern times. It is expected from the designers to build
efficient and dynamic production facilities in order to get the best potential results in terms
of profitability, throughput, and lead time due to international competition. It is imperative
to study the performance measurements in the field of tile production designs in order to
improve the efficiency and to optimize the processes. Simulating the performance of a real
system without making any changes to the real system is one approach for studying its
performance. It is extremely likely that the behavior of the real system can be observed and
understood via simulation, since the simulations are cost effective and can help to forecast
the resulting patterns. The recurring geometrical form in a design pattern has no limits
because the result is a series of independent geometrical forms that are more or less close to
each other. When it comes to tile designs, the repetitive form necessitates a specific shape
in order to minimize gaps or overlapping. Any figure, picture, or drawing utilized as a unit
motif in a pattern design is referred to as a “geometrical form" [15].
During this research, the numerical model of the tiling design is simulated using
MATLABTM software, the main goal of simulation in tiling pattern or tiling design is to
facilitate decision-making so that desired goals can be met quickly. To enhance the system,
simulation tools can produce better results that indicate a significant gain in productivity,
cycle time, and throughput time. In the literature, evidence is available, where the numeri-
cal simulations have been built to analyze the behavior of complex patterns and to simplify
 September 26, 2022 8:5 0218-348X

the nonlinear mathematical models. The numerical simulations help the reader to better
interpret and utilize the research findings [16, 17, 18, 19].
In this manuscript, we have provided the step by step description of the method with
numerical algorithms and innovative tile structures. The statistical analysis of the resulting
images and patterns is conducted with the aid of machine learning regression and classifi-
cation tools. In the next section, the methods are illustrated with the flow chart. Next useful
results and innovative conclusions are drawn.

2. Materials and Methods

2.1. Materials

In 1975, the mathematician Benoit Mandelbrot coined the term “fractal" from the Latin word
fractus, which means irregular or fragmented, to describe objects that were too irregular to
fit into a traditional geometrical setting. Fractal is a new type of subject that has emerged
in the last 40 years. The mathematical concept of fractal is difficult to classify formally,
even for mathematicians, but key features can be understood with a little mathematical
background. These erratic and fragmented shapes can be found almost in every architecture,
designed with innovative approaches. Fractals, at their most basic, are a graphic illustration
of a repeating pattern or formula that begins simple and becomes increasingly complex.
Fractals are broadly classified into two types: random fractals (or non-deterministic) and
deterministic fractals.
The random fractals basically lack the self-similarity, these are not similar to their system-
atic counterparts, their non-uniform appearance is often more akin to natural phenomena
and is the preferred method when modeling coastlines, topographical surfaces, or cloud
boundaries, is the subject. However, random fractal implies that the rules for generating
fractal graphics remain unchanged. So, while the graphics produced by two operations at
different times may have the same fractal dimension, the final generated graphics will differ
due to the influence of random factors, inferring that random fractals are not repeatable.
Among the fractal definitions available in the literature, the approach that is used is the
Falconer’s, which, rather than a computationally efficient definition, provides a concise
indication in the form of a list of properties. A fractal, G is taken to be a collection of objects
that exhibit many of the following characteristics:
September 26, 2022 8:5 0218-348X

• The structure must be fine and recursive.

• The traditional geometric language is not sufficient to define G.
• There is self-similarity (statistical) in G.
• Dimension of G is greater than its topological dimension. recursively.

2.2. Methods

2.2.1. Fractal Patterns Algorithms

The fractal patterns can be generated using analytic models as well as by virtue of numerical
simulations, using simple and parallel pool algorithms [20]. Different softwares provide
different tools to generate fractals.
Wang and Zuo [21] generated grid map from the sampling data, converted map from
space to Fourier domain, with the aid of fast Fourier transformation and obtained the power
spectrum. Their algorithm then followed the process of segmentation and filtration. Finally
inverse Fourier approach was utilized to generate the final results.
Similarly, Lu et’al. [22] discussed in detail the algorithms available for Fractal generation
for the architecture design with the aid of smart programming tools. Several other studies
were reported during the past decade to support the importance of algorithms and their
smart design in generating fractals [1].
One of the leading softwares, Matlab TM provides the users very swift algorithms
for Fractal generation. These include Fractal-Toolkit, Hausdorff-Fractal-Dimension tool,
Minkowski’s loop fractal toolbox, Koch-curve-fractal tool box, 2D-Lyapunov fractal,
Icosahedral-fractal, Icosahedral-fractal and many more. An important thing to keep in
view, before utilizing these Fractal tool boxes is that most of these algorithms require good
processors and parallel processing.
In the next section we will outline the properties of Fractal patterns and the algorithmic
approach, used during this research, to generate tile patterns. The algorithm designed
during this research is comparatively cost effective than the other algorithms outlined
above and is thus user friendly.
 September 26, 2022 8:5 0218-348X

2.2.2. Fractal Patterns Application

Inspired from the available algorithms for fractal generation [23], during this research, a
numerical algorithm is used that is looped to generate four different patterns, in 16 different
orientations. These patterns are defined as the Julia-Fractal, Mandelbrot-Fractal, Newton-
Fractals and Nova-Fractal respectively.
The four approaches that we have used while designing the algorithm, can be better
interpreted with the aid of the following schematic 1. The fractal graphic design program
starts with the master node, where the master node transmits work to the next node as
shown in the frame diagram. The key node will then receive the row number and start
working on the classification and derivation of fractal art graphics, it will then start further
processing on the measurement scale range of fractal art graphics.
To proceed, the algorithm will require the input for the expected designs based on
the requirement of the architect, and the algorithm will select the preferable technique
accordingly. To begin the process, we choose one of the four fractal designs from our four
options. The node passes the task to tile engineering design, to process pattern coloring
and display the fractal art when the measurement scale of fractal art graphics is finally
simulated. If the display design does not meet the expected outcomes, the process will
mutate and restart at the master node. However, if the outcome is satisfactory, the algorithm
will proceed to interpret and transmit the image. At that point, the algorithm will stop and
exit the program.
September 26, 2022 8:5 0218-348X

Figure 1: The classification framework of innovative fractal art graphics of tiles printing,
based on the needs of the consumer.

3. Results and Discussion

In this section, we will discuss the numerical results generated with the aid of four ap-
proaches.
 September 26, 2022 8:5 0218-348X

3.1. Julia Fractals Graphic

The Julia set and the Fatou set are two complementary sets (Julia “laces" and Fatou “dusts")
defined from a function in the context of complex dynamics.
The Julia set of a function f is commonly denoted J(f), and the Fatou set is denoted
by F(f). Julia sets are a great example of how a simple process can result in wonderfully
intricate sets. Julia sets are born from the iteration of functions in the complex plane and are
named after the French mathematician Gaston Julia. They are based in the field of complex
analysis [24].
theoremet f : C → C be a polynomial of degree n ≥ 2 with complex coefficients,
f(z) = b0 + b1 z + b2 z2 + . . . + bn zn where bi ∈ C, bn , 0. To start looking at Julia sets,
we’ll need the following:
L. • Let f k (w) = f(f(. . . (w) . . .)) be the k-fold composition of f with itself.
• If f(w) = w for some w ∈ C, we call w a fixed point of f.
• If f p (w) = w for some integer p ≥ 1, w ∈ C, we call w a periodic point of f.
• If w is a periodic point of f, the least such p ∈ N≥1 such that f p (w) = w is called the
period of w.
• If w is a periodic point with period p, then f (w), . . . , f p (w) is the period p orbit of w.
′
• If w is a periodic point with period p, and ( f p ) w = λ, where prime denotes the
complex differentiation.

super attractive periodic point of t if |λ| = 0

attractive periodic point of t if 0 ≤ |λ| < 1
indifferent periodic point of f if |λ| = 1
repelling periodic point of t if |λ| > 1.

• The Julia set J(f) of f may be defined as the closure of the set of repelling periodic
points of f.
• The Fatou, or stable, set F(f) = C/J(f) of f is the complement of the Julia set of f.

Based on the algorithm (figure 1), the following patterns (figure 2) are generated with
the aid of MatlabTM framework.
Julia fractals caught a lot of attention because the graphics are spectacular. This Julia
September 26, 2022 8:5 0218-348X

fractal tiling is suitable for building decoration and interior design. The fractal pattern
design presented in Julia fractal images (figure 3) has some advantages such as it has the
ability to effectively reflect the characteristics of real scenery and texture. The contrast of
white and black (light and darkness, day and night) has a long history of metaphorical use.
We present a revolutionary pattern design technique combining fractal geometry and visual
texture generation. These Julia fractal structures apply to manipulate texture generation
with light colors, where the shade of the pattern could be easily changed. Experiments
revealed that patterns with various styles can provide multiple forms to decorate the
environment effectively while using the new tile designs.

((a)) f (z) = z2 + c, c=-0.28+0.008i ((b)) f (z) = z2 + c, c=-0.79+0.15i

Figure 2: Different orientations of Julila fractals.
 September 26, 2022 8:5 0218-348X

Figure 3: A wall visualizations with fractal tile pattern using Julia.

3.2. Mandelbrot Fractals Graphic

The Mandelbrot set is made up of iterations in the complex plane of specific quadratic
functions. It arises from the study of Julia sets of functions of the form fc (z) = z2 + c; those
which we have briefly studied in the previous section.
theoreme define the Mandelbrot set to be the set of parameters c for which the Julia set
of J(fc ) is connected.
W . M = c ∈ C : J(fc ) is connected.

The Julia set and the Mandelbrot set have a significant difference in terms of iteration.
We must iterate z always starting from 0 and varying the value of c in the Mandelbrot set.
The Julia set iterates for a constant value of c and a variable value of z. In other words,
the Mandelbrot set is in the parameter space, or c-plane, whereas the Julia set is in the
dynamical space, or z-plane.

Following the same procedure explained in section 2.2, figure 4 provides interesting
patterns.
We can see that these gloomy tiles are darker in color and are typically used in more
professional environments, such as offices. Dark diagonal-patterned tiles in small kitchens
deceive the eye into believing the area is larger than it is. Dark flooring can produce a
rich contrast or an unexpected impression in a well-appointed lavatory, as shown in these
September 26, 2022 8:5 0218-348X

Mandelbrot images (see figure 5), respectively). Because they are easy to clean and do not
retain dirt and dust for longer duration, they can be used in restrooms, desk areas, offices,
and restaurants. These images can be used for a variety of dark tile surfaces, including
glazed finish, glossy finish, matte finish, and textured finish.

((a)) f (z) = z4 + c, c=-0.6+0i ((b)) f (z) = z5 + c, c=-0.6+0i

Figure 4: Two different presentations of Mandelbrot graphs.

Figure 5: A wall visualization with fractal tile pattern using Mandelbrot.

 September 26, 2022 8:5 0218-348X

3.3. Newton Fractals Graphic

“Isaac Newton’s method" was used to create Newton Fractals. They have a completely
unique appearance compared to most fractal sets. The existence of Newton fractals is proof
that Isaac Newton was a brilliant mathematician in his time.

These fractals are ideally suited for bathrooms, kitchens, and dining rooms. Because
bright colors create the illusion of light. For a strong, saturated mix of shades. The fractal
patterns in figures 6 and 7 gives a different impression. Using Newton fractal art with
vibrant colors may create a rich contrast and draw attention to generate tiles for a stunning
floor with a constantly repeated design. These patterns have high design parameters, a big
occupation area and a sophisticated design process.

((a)) The basins of attraction, f(z)= z3 +1, c=- ((b)) Magnification of the middle, f(z)=z3 +1,
0.5-0.8660254i c=-0.5-0.8660254i
Figure 6: Different magnifications for Newton’s fractals.
September 26, 2022 8:5 0218-348X

Figure 7: A wall visualization with fractal tile pattern using Newton.

3.4. Nova Fractals Graphic

Nova is really a large class of fractals since any function f(z) can be used. Varying ‘R’ changes
the shape of the fractal. The classic Nova fractal uses f(z) = zp -1, and therefore produces the
following equation that is iterated:
p
f (zn )
zn+1 = zn − R ′ p−1 +c
f (zn )

Mostly images use p=3, which represents the search for the cubic roots of unity, and yields:

(z3n )
zn+1 = zn − R 3(z2n )
+c

The new designs in tile construction can be driven by the creative usage of Nova fractal
patterns in various shapes, textures, and patterns. The numerical results generated by using
Nova present the fractal patterns in a more artistic way, matte colors in these tile textures,
offer a softer, warmer impression as we can see from figures 8 and 9 respectively. These
encaustic tiles and imitations provide a Victorian style. They can be used in many types of
 September 26, 2022 8:5 0218-348X

projects, from kitchen back splashes, floors, outdoor spaces and decorative wall features.

((b)) p(z) = z3 − 1, c = 12 eiθ , θ ranges from 0

((a)) p(z) = z3 − 1, c ranges from -1 to 1 to 2π
Figure 8: Comparison of two designs with different ranges.

Figure 9: Wall visualization with fractal tile pattern using Nova.

September 26, 2022 8:5 0218-348X

3.4.1. SVM-Classification

The fractal patterns, designs and resulting images generated during this research were
processed using the image segmentation and filtration. The resulting statistical data was
classified using the state of the art classification tools [25], with least loss of important
features of the patterns. The judgment/ selection criteria was defined according to the user’s
choice, features (broad/ narrow, recursive, size, shape), architecture type and colour contrast
respectively. The classification results from the support vector machine learning classifier
are presented in figure 10. These results were generated following the step (1) the fractal
patterns with different orientations are used to produce different images. These images
were segmented and the statistical data, including the entropy, contrast and homogeneity
was processed and classified. The two response variable was taken to be the type of illusion
that the images made, broad or narrow. The results obtained showed 84.3% accuracy.

Figure 10: Classification of fractal patterns, according to wall broad or narrow illusion.

A summary of all the algorithms is provided in table 1.

 September 26, 2022 8:5 0218-348X

Method Iterative Description Limitation Characterization

Julia c non-variable infinite dynamical Plane
Mandelbrot variable c infinite parametric Space
Newton class of fractal infinite polymetric space
Nova variable c infinite polymetric space
Table 1: Comparison of the algorithms.

4. Conclusions

Integrated fractals in a tiling design with geometric repetition provide creative patterns
with continuity. During this research, a unique iterative technique is used to create several
classical fractal patterns that combine multiple characteristics and different orientations.
The creation of these methods by using MATLABTM with the repetition principle of fractals
provides the production of various forms with different complexities. Innovative fractal
patterns are created and tested for accuracy and perfection, when used in architecture. In
order to carry out the essential research, fractal art on tiles was combined with different
pattern styles, keeping in view the applications of fractal art with smart tools. The discovery
of Fractal Geometry has made it possible to automatically find thousands of attractive tile
themes, and designers can now run large numbers of simulations without limitations using
a digital system. Fractals can be more frequently and swiftly employed in the future design
sector, and in the field of product design to generate a higher artistic level.

Compliance with Ethical Standards:

Acknowledgments:

Not applicable.

Declaration

The authors declare that there is no conflict of interest.

References

1. Wang W., Zhang G., Yang L. and Wang W., Research on garment pattern design based
on fractal graphics, EURASIP Journal on Image and Video Processing 2019(1):1–15, 2019.
2. Yu Z., Sohail A., Nofalc T. A. and Tavaresd J. M. R., Explainability of neural network
clustering in interpreting the covid-19 emergency data, Fractals , 2021.
September 26, 2022 8:5 0218-348X

3. Yu Z., Arif R., Fahmy M. A. and Sohail A., Self organizing maps for the parametric
analysis of covid-19 seirs delayed model, Chaos, Solitons & Fractals 150:111202, 2021.
4. Wang F., Sohail A., Tang Q. and Li Z., Impact of fractals emerging from the fitness
activities on the retail of smart wearable devices, Fractals , 2021.
5. Yu Z., Sohail A., Arif R., Nutini A., Nofal T. A. and Tunc S., Modeling the crossover be-
havior of the bacterial infection with the covid-19 epidemics, Results in Physics 39:105774,
2022.
6. Sohail A., Yu Z., Arif R., Nutini A. and Nofal T. A., Piecewise differentiation of the
fractional order car-t cells-sars-2 virus model, Results in Physics 33:105046, 2022.
7. Chen Z., Application of iteration function system for ceramic tile design, 2020 IEEE
Intl Conf on Parallel & Distributed Processing with Applications, Big Data & Cloud
Computing, Sustainable Computing & Communications, Social Computing & Networking
(ISPA/BDCloud/SocialCom/SustainCom), IEEE, pp. 1475–1480, 2020.
8. Batty M., Kay D., Healey P., Dunn R. and King A., Reviews: The Beauty of Fractals: Im-
ages of Complex Dynamical Systems, Computer Models in Environmental Planning, Research
in Planning and Design 14. Critical Rationalism and Planning Methodology: Conflict Man-
agement in Urban Planning: Urban Innovation: The Transformation of London’s Docklands
1968–84, Qualitative Modelling of Complex Systems: An Introduction to Loop Analyses and
Time Averaging, a Concrete Atlantis: US Industrial Building and European Modern Architec-
ture 1900–1925, 1987.
9. Watt A., Fundamentals of three-dimensional computer graphics, Addison-Wesley Longman
Publishing Co., Inc., 1990.
10. Agirbas A., A proposal for the use of fractal geometry algorithmically in tiling design ,
2019.
11. Mampusti M. A., Software to produce fractal tilings of the plane , 2014.
12. Lin X. and Liu W., The application of fractal art in ceramic product design, IOP Conference
Series: Materials Science and Engineering, IOP Publishing, p. 012003, 2019.
13. Lefler S., Fractals and the chaos game, MAT Exam Expository Papers p. 24, 2006.
14. Ouyang P., Tang X., Chung K. and Yu T., Fractal tilings based on successive adjacent
substitution rule, Complexity 2018, 2018.
15. Nadir U., Habib T., Khattak S. B. and Noor I., Performance improvement using simula-
 September 26, 2022 8:5 0218-348X

tion tool in a tiles production facility, Mehran University Research Journal Of Engineering
& Technology 39(1):117–132, 2020.
16. Sohail A. and Nutini A., Forecasting the timeframe of coronavirus and human cells
interaction with reverse engineering, Progress in Biophysics and Molecular Biology , 2020.
17. Sohail A., Yu Z., Arif R., Nutini A. and Nofal T. A., Piecewise differentiation of the
fractional order car-t cells-sars-2 virus model, Results in Physics p. 105046, 2021.
18. Yu Z., Abdel-Salam A.-S. G., Sohail A. and Alam F., Forecasting the impact of environ-
mental stresses on the frequent waves of covid19, Nonlinear Dynamics 106(2):1509–1523,
2021.
19. Yu Z., Ellahi R., Nutini A., Sohail A. and Sait S. M., Modeling and simulations of
covid-19 molecular mechanism induced by cytokines storm during sars-cov2 infection,
Journal of Molecular Liquids 327:114863, 2021.
20. Kamer Y. and Hiemer S., Data-driven spatial b value estimation with applications
to california seismicity: To b or not to b, Journal of Geophysical Research: Solid Earth
120(7):5191–5214, 2015.
21. Wang J. and Zuo R., A matlab-based program for processing geochemical data using
fractal/multifractal modeling, Earth Science Informatics 8(4):937–947, 2015.
22. Lu X., Clements-Croome D. and Viljanen M., Fractal geometry and architecture design:
case study review, Chaotic Modeling and Simulation (CMSIM) 2:311–322, 2012.
23. Pawar
A., Mandelbrot set and julia set , matlab central file exchange. retrieved september
26, 2022., Mathworks (https://www.mathworks.com/matlabcentral/fileexchange/24740-
mandelbrot-set-and-julia-set).
24. Cooper F., Introduction to fractals and julia sets .
25. Sohail A., Fahmy M. A. and Khan U. A., Xai hybrid multi-staged algorithm for routine
& quantum boosted oncological medical imaging, Computational Particle Mechanics pp.
1–11, 2022.