DOI: https://doi.org/10.

51846/vol5iss4pp68-74 Pakistan Journal of Engineering and Technology, PakJET

ISSN (p): 2664-2042, ISSN (e): 2664-2050
Volume: 5, Number: 4, Pages: 68- 74, Year: 2022

Image Processing Based Pattern Recognition and

Computerized Embroidery Machine
Aisha Shariq 1, Afshan Khan 2, Abdul Moid Khan 3, Muhammad Khurram 4, Muhammad Farooq Umer 5
and Muhammad Salik Salam 5
1
Department of Computer Science, Sir Syed University of Engineering & Technology, Karachi, Pakistan
2
Department of Computer Science, Nazeer Hussain University Karachi, Pakistan
3
Department of Electronic Engineering, Sir Syed University of Engineering & Technology, Karachi, Pakistan
4
University of Technology and Applied Sciences, Nizwa Sultanate of OMAN
5
Department of Computer Science, University of Alabama at Birmingham, 1402 10th Avenue S. Birmingham, AL 35294, USA
Corresponding author: Aisha Shariq (e-mail: [email protected]).
Received: 8-09-2022 Revised: 19-11-2022 Accepted: 17-12-2022
Abstract- In the last decades, embroidery getting much attention in industries and academics. Although the technique is very old,
it is still very much adopted. The current era of machine learning and smart devices makes this method even more attractive,
easy to adopt and high accuracy. The proposed research work aims to highlight the implementation of new smart devices in this
field, and the work is based on microcontrollers. Adopting the proposed work will be very easy and feasible to transform from
the old sewing machine to a new smart method. Using smart devices, it is very easy and precise to get the embroidery fulfilling
the user's demands. It gives simple raster images, which are used to convert into digital images to portray any new product. The
current work is a novel idea of generating simple images rather than complicated ones, which are difficult to draw. In our
method, we introduce the digitization process to input any image file used for vectorizing raster images, where vector images
can easily convert into a DST embroidery file, a commonly used embroidery format originally from Tajima. It also has a Gcode
file. The machine runs on a Gcode file which can be easily generated and runs on a product.

Index Terms-- Intelligent embroidery, image processing, pattern design, auto-digitizing, CNC machine, rasterizing, vectorization,
transformation.

I. INTRODUCTION textile machines is a problem for Pakistan's textile sector. We

In Pakistan, embroidery is regarded as one of the known may reduce our import ratio by producing this machine,
traditional crafts. Embroidery is a sort of art applied to fabric or progressively improving our GDP and inflation rates. Villages
various types of clothes and accessories using needles, thread, are known for their gorgeous textiles, which include Ralli quilts.
or yarn. In Pakistan, needlework is done on a big scale and is It's a way of repurposing cloth. Older fabrics are sewn together,
highly traditional, with most of the workers, both men, and and then some embroidery is done with that material to mask its
women, coming from villages. This is how most village flaws. Khairpur Mirs is well known for rally-making, and each
households generate money nowadays. Many domestic item manufactured here sells for at least one to two lakes. A
products, such as badges, flags, bed sheets, towels, uniforms, single bedsheet can take up to a week to make, while a double
pouches, handkerchiefs, slippers, shoes, and the most popular bedsheet might take up to a month.
work, Rally, are to be embroidered in Pakistan. People This rally-making activity is viewed as a source of money for
throughout the globe appreciate beautiful needlework work. them and a source of cash for our country. By producing many
However, the market demand cannot be met due to inadequate pieces daily, this machine may speed up its job, producing
manufacturing efficiency. Only in refined society did revenue for the government. Rural women must demonstrate
elaborately embroidered garments, religious artifacts, and their abilities and endeavours to promote entrepreneurship and
domestic things become popular [1]. skill development among their peers. The country's
The textile industry is Pakistan's most important, accounting unemployment rate will fall. Our inflation rate and GDP will
for more than 67 percent of overall export revenues and 46 continue both rises at the same time. Many textiles-related
percent of manufacturing [2]. As the industrial revolution initiatives will be completed in the allotted period [3]. Today's
advanced, machine embroidery and mass production expanded. fashion design students have unique design concepts and can
Machine embroidery may be accomplished in several forecast future season trends rather than squandering many
techniques. Pakistan has imported more than $4 billion of textile hours designing and bringing to life those details. They only
and garment machinery in recent years. The low productivity of need to utilize the machine tools and may quickly sketch their
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
68
inventive concepts, allowing them to establish their firm or Our research's main process is to turn an old sewing machine into
expand their sector. an embroidery machine with user-interactive software that
The demand for automated machines has skyrocketed in accepts various input files and converts them to a computerized
recent years, thanks to the introduction of microprocessors that embroidery file. Create and update an SVG image and transform
enable automatic machines to do numerous settings-related it into a digital embroidery file compatible with the machine [5].
activities efficiently. Because of the user interface's reliability, Upload the proper format file via USB, Bluetooth, or Wi-Fi, and
it's much easier to learn how to use the software and develop the place the appropriate hoop. Fabric stabilizing, placing hoops,
design. This computer contains pre-installed designs in its placing and changing the ribbon, take-up lever, oiling, and bobbin
memory and a Wi-Fi connection for downloading additional winding are some of the works to be done manually [6]. Among
pictures. It will accept all the user's inputs and begin creating the these, the embroidery and frame movement is the most important
design automatically. There are already built-in audio and video since they finished the embroidery independently and
tutorials for running this software. Modern stitching machines harmoniously. The wheel pushes the needle and thread in the Z
are adaptable to various automated stitching procedures, saving direction to move up and down, and the frame reciprocates in the
time, enhancing output, and improving efficiency. Because of X and Y directions to produce a beautiful embroidered work.
its continuous technological advancement, the contemporary
sewing machine is quite simple to operate.

II. BASIC WORKFLOW OF THE DESIGN AND

MANUFACTURE
EmbroidAI aims to be a proficient embroidery digitizing platform
based entirely on free, open-source modules, able to be used on
different types of computers or with different software packages.
Our goal is to be accessible for hobbyists and digitizers with the
powerful auto-digitizing tool, which helps in saving time, effort,
and money. To create and update a beautiful embroidery design,
easily share with others, and enjoy the auto-digitizing tool. The
design shown in Fig. 1 delivers gorgeous embroidery designs
with ease and confidence [4]. Manage the machine and all its
components to work synchronously. This software aims to create
easily, update, visualize, save, and share digital embroidery files.
Users can insert the file via USB, share it by Bluetooth, and
upload the embroidery file through the accessible internet.
This research presents a computerized embroidery machine based
on the Arduino UNO controller. An old sewing machine can be
converted into an embroidery machine with some hardware
components, including a 2dplotter, stepper motor, stepper driver,
controller, CNC shield, and power supply. To change the
platform of an embroidery machine according to the embroidery
design selected by the user. A user may upload a design, such as a
simple vector graphic, and have it transformed into a digitized
image, which can then be used to create personalized embroidery
that can be applied to a product.
This paper provides the most effective ways of tackling our
research's most demanding task: auto-digitizing utilizing various
AI technologies. This input file goes through several steps,
including noise reduction, background removal, conversion to
black and white, enhanced image resolution, edge detection, and
contouring, before being rendered accurately into an SVG vector
image, which can be edited, and then transformed into our
machine-readable file, which converts directly to Gcode. We've FIGURE 1: Basic Workflow Design
discussed approaches for producing high-quality results in real
time. Strong animation tools for embroidery are typical textile III. LITERATURE REVIEW
embellishments, and producing and drawing them automatically
is useful. This study demonstrated the best conversion tool for A. IMAGE PROCESSING
converting any embroidery file into a machine-readable format. Image processing aims to increase the image quality so that it can
be easily comprehended by people or robots [7]. Its methods
69
transform input images into high-quality output images. de-noising performance, CNN-based noise reduction models have
Continuous pictures and discrete images are the two types of been added. However, these approaches must be trained on a big
images. Optical devices that receive analog signals, such as the dataset better to understand the characteristics of pictures for
human eye and analog cameras, produce continuous pictures. noise reduction. Furthermore, the research found that while the
Continuous pictures are converted into discrete images through deep CNN model learns effectively on many samples when labels
the digitization process. Digital cameras and scanners, for are randomized, it may overfit the samples [13].
example, are equipped with digitization features that allow them Auto tracing is a technique for converting a bit-mapped picture
to generate distinct pictures [8]. into a vector image. Grain, halftone dots, and other restrictions
Artificial intelligence (AI) is a machine's capacity to perform may exist in bit-mapped pictures, which are represented as dots.
activities normally performed by people and needs human Converting them to vector images can address this problem. The
intelligence and judgment, both of which Python excels at. It is open-source tools Map traces, Pypotrace, and auto trace display
perfectly suited for usage in this artificial intelligence sector, and picture files to recognize paths, lines, and objects [14]. When you
its popularity will continue to grow [9]. The most extensively have a raster graphic that is too tiny for your purposes, you may
utilized AI scripting language is Python. The library is an use an auto-digitizing tool to convert it to a vector. It's a quick and
interface, or a bundle of modules, offered by multiple sources easy procedure that yields amazing results. Improved picture
such as PyPi, which includes an open-source code that allows resolution for resizing and detecting image file edges [15]. Since
users to do specific tasks or access certain functionality. then, it has evolved due to the most advanced algorithms and high
We need to determine borders and properly mark regions from computation power (see Fig. 2). Humans can now create facial
the scanned input files to analyze image file embroidery designs. recognition software that is as accurate as humans [16].
In this paper, we provide the most effective ways of tackling our
research's most demanding task: auto-digitizing utilizing a variety
of AI technologies. This input file goes through several steps,
including noise reduction, background extraction, small object
removal, finding shapes, conversion to a binary image, enhanced
image resolution, edge detection, and contouring, before being
rendered accurately into an SVG vector image, which can be
edited, and then transformed into our machine-readable file,
which converts directly to Gcode [10].
Canny edge detection comes to mind as the initial method for
extracting edges. On the other hand, Canny edge detection looks FIGURE 2: Imager processing input and output image
for sudden changes in image brightness in an image. An
embroidered design has well-defined boundaries and is not a large B. MODELING
picture. The essential task is to extract areas and associated In this research, there are four modules to be covered to create a
borders. The borders of an embroidered pattern are always beautiful embroidery design. We have done many processes to
surrounded by pixels, even if the scan appears black and white. allow several types of files to be input, including raster, vector
As a result, we begin by creating a true binary image with basic. images, or any embroidery files.
The next stage is to locate related chains of edges, which may be
used to locate boundaries and regions [11]. a) DST Embroidery File:
The first module to be stitched is the Data Stitch Tajima
The Hough transform is a module used for identifying forms in
embroidery file, a widely used commercial embroidery file
photographs. Lines, circles, and ellipses have all been extracted
developed by Tajima [17]. This is our machine-readable file,
using this method. Converting an image from Cartesian to polar
which automatically translates to Gcode. Stitch and Jump are
coordinates is a common feature extraction approach. In the the only commands that it understands. After finishing one
Hough space, any point in the picture space is represented by a colour, it switches to the next. We may use the pause command
sinusoidal curve. Even though this model form is simple, it to replace the ribbon and bobbin in the interim. Finally, our
becomes extremely problematic when dealing with complex lovely design will be stitched.
forms owing to noise and shape imperfection and the challenge of
identifying vertical line slopes [12]. The CHT solves this b) Embroidery File Converter:
difficulty by inserting a transformation of the shape's centroid in In this platform, we provided the best conversion tool to insert
the parameter space's x-y plane. any embroidery file type, which is the second module of our
Noise removal, enhanced image resolution, contouring, edge, research. We've demonstrated how to convert any embroidery
and contrast enhancement are learned simultaneously for effective file into a machine-readable format using the greatest
performance improvement without much training data. Applying conversion tool available, which can then easily be translated to
LP decomposition on de-noised photos prevents colour distortion. Gcode. Our machine moves in three directions, using x-y
Using spatially correlated information, edge information is commands to move back and forth and z commands to move up
augmented to generate high-quality output pictures. To improve and down [17].
c) Raster Image:
70
The raster image is the fourth and last component of our
research. This paper provides the most effective ways of
tackling our research's most demanding task: auto-digitizing IV. METHODOLOGY
utilizing various AI technologies. This input file goes through A. Digitization
several steps, including noise reduction, background removal, Before digitizing, analyze and plan the order of the pattern
conversion to black and white, enhanced image resolution, edge design forms and embroidered stitches. Automatic digitization
detection, and contouring, before being rendered accurately into can turn a basic form or embroidered item into a pattern design.
an SVG vector image, which can be edited, and then For developing various forms or pattern design components,
transformed into our machine-readable file, which converts many tools are available. After selecting the proper tool, digitize
directly to Gcode. & digital technology by designating reference points along the
d) Vector Image: contour.
A vector image is a computer picture used in graphic design that B. Professional Digital Technology
typically includes bright, striking colours, such as logos and This application will provide you with specialized digital skills
printed images. It may be saved in EPS, SVG, and AI. We have that will allow you to save time while creating special effects
mostly worked on a Scalable Vector Graphic (SVG) picture in and machine functionality. To provide input tools including
this research. Our SVG Draw Bot provides best practices for digital mosaic stitching, stars, and branches. Contouring,
delivering high-quality outputs in real-time and robust embroidery, and contour deepening are all methods that may be
animation features for automatically transforming vector image used. Use the application to digitize embroidered or sequined
files into editable forms and drawing vector images using pen towels. The embroideries are removed from the pattern and
and pencil tools. This input file may be turned into a machine- exported as separate files or sent to the cutter. Large areas can
readable file, which can then be translated to Gcode. also be embroidered with low-stitch cross stitches.
Table I The SVG sketch commands included in this work, their C. Computer-Aided Design
parameters and a visualization. The beginning (x1, y1) is Linear, banded, and regional input approaches were used for the
generally specified as the previous command's terminal. computer embroidery. In consequence, digital technology and
editing modules have been merged with advanced graphics
TABLE I: processing technology, adequately enhancing the software's
SVG SKETCH COMMANDS
efficiency to create, update, process, and manage pattern
designs based on creative embroidery knowledge, an automatic
Command Arguments Description Visualization
needle-filling method, and a repository that can make any
<SOS> ∅ ‘Start of SVG’
drawing accurate [19]. To detect and fill the appearance of a
token broken line and boost design, and production efficiency, use the
automated single-needle linear input and single manual needle
M x2, y2 Move the based on the side of the input, bilaterally symmetrical input, and
cursor to the medial input [20].
(Move To) endpoint (x2, D. Drawings Sewn with Intelligence
y2) without
drawing There are plenty of methods, such as pattern preparation, to
anything. control the embroidery machine and embroidered hoop to
perform multiple operations. A pattern master created this
L x2, y2 Draw a line to process. The phrase derives from the mechanical embroidery
the point (x2, machine, which records stitches on punched tape. Nowadays,
(Line To) y2).
embroidery machines can recognize and perform their actions,
using a mechanical or more modern way to record in memory
C qx1, qy1 qx2, Draw a cubic for those who need to create a pattern.
qy2 x2, y2 Bezier curve
with control Three types of tools are used in a computerized embroidery
(Cubic
Bezier) points (qx1, machine, embroidered plate production software, and a U disc.
qy1), (qx2, qy2) The basic steps for creating and modifying embroidery design
and endpoint using a computerized embroidery machine are as under below:
(x2, y2).
1. The initial process is the design of embroidered
z ∅ Close the path designs, also known as plate makers. The component
by moving the of the staff's job involves plate manufacture. Designers
(Close Path) cursor back to may easily generate patterns using this tool, which can
the path’s
starting position subsequently be stored on a computer's hard drive.
(x0, y0). 2. Locate and open the embroidery pattern file in the
following steps. The suffix DST files are being altered.
<EOS> ∅ ’End of SVG’ 3. The third component is to choose or upload the file into
token.
the computer embroidery machine, modify it, and save

71
 it into memory using the control buttons on the Our machine moves in three directions, like a 3D printer, using
machine's panel. x-y commands to move back and forth and z commands to
4. Now place the frame of the selected size. After sliding move up and down [24].
the embroidery frame up and down until it is in a
perfect position, click the embroidery confirmation
button. This will establish the embroidery's
provenience.
5. Choose an embroidery colour to begin embroidering.
6. During the actual manufacturing process, several
factors (such as a scarcity of needles and thread) result
in the loss of part of the stitch. Use the features of
broken line detection, back, and re-embroidered to
finish the embroidery task (see Fig. 3).

FIGURE 4: DST Embroidery File

FIGURE 3: Machining Processes

V. SIMULATION RESULTS

Simulation is the process of simulating the function of a

meaningful process or system through time. For simulations,
FIGURE 5: Brother Embroidery File
designs are essential; the prototype represents a major trait or
actions of the selected framework, whereas the model depicts The third component is a Scalable Vector Graphic (SVG) image
the evolution over time. Four works are covered in this section. (see Fig. 6). Clean lines, flat colours, and uniform sterility are
The first component to be stitched is the Data Stitch Tajima all hallmarks of vector drawings. This file is scalable and may
embroidery file (see Fig. 4) [16]. This is our machine-readable be viewed from any location at any time. It's easy to alter and
file, which automatically translates to Gcode. Stitch and Jump resize images to any extent beyond the drop of detail. Our SVG
are the only commands that it understands. After finishing one Draw Bot provides best practices for delivering high-quality
colour, it switches to the next. We may use the pause command outputs in real-time and robust animation features for
to replace the ribbon and bobbin in the interim. Finally, our automatically transforming vector image files into editable
desired design will be stitched [21]. Fabric is one of the vintage forms and drawing vector images using pen and pencil tools.
inventions of human beings [22]. In recent years, fabric pattern This input file may be turned into a machine-readable file,
(texture) recognition has gained much attention and made great which can then be translated to Gcode.
achievements [23].
The second element to be stitched is the Brother Embroidery file
(see Fig. 5). Our machine accepts multiple formats of
embroidery files and transforms them into a machine-
compatible format with the powerful conversion tool. Using the
greatest conversion tool, we've demonstrated how to convert
any embroidery file into a machine-readable format. This input
file may be turned into a machine-readable file, which can then
be translated to Gcode.

72
 The authors declare they have no conflicts of interest to report
regarding the present study.

CONFLICT OF INTEREST
The Authors declare that they have no conflicts of interest to
report regarding the present study.

REFERENCES

[1] [Onine] “5 Traditional Embroideries of Pakistan that Make a Serious

Style Statement,” YourLibaas.
https://www.yourlibaas.com/blogs/fashion/5-traditional-embroideries-of-
pakistan-that-make-a-serious-style-statement (accessed Mar. 24, 2022).
[2] [Online] “Textile Fabric.pdf.” Accessed: Mar. 24, 2022. [Online].
FIGURE 6: SVG Image Available: https://www.sbp.org.pk/departments/ihfd/Sub-
Segment%20Booklets/Textile%20Fabric.pdf
The PNG raster image is the fourth and last component of our [3] [Online] “PES_2020_21.pdf.” Accessed: Mar. 24, 2022. [Online].
Available: https://www.pc.gov.pk/uploads/cpec/PES_2020_21.pdf
research. The raster image (see Fig. 7) is a rectangular grid of [4] M. Sattler, R. Sarlette, and R. Klein, “Per Christensen and Daniel Cohen-
pixels, which is portable and easy to share. This paper provides Or (Editors) Efficient and Realistic Visualization of Cloth.”
the most effective ways of tackling our research's most [5] J. Kim and F. Pellacini, “Jigsaw image mosaics,” in Proceedings of the
demanding task: auto-digitizing utilizing various AI 29th annual conference on Computer graphics and interactive
techniques, New York, NY, USA, Jul. 2002, pp. 657–664. DOI:
technologies. This input file goes through several steps, 10.1145/566570.566633.
including noise reduction, background removal, conversion to [6] A. Hertzmann and D. Zorin, “Illustrating smooth surfaces,” in
black and white, enhanced image resolution [14], edge Proceedings of the 27th annual conference on Computer graphics and
detection, and contouring, before being rendered accurately into interactive techniques, USA, Jul. 2000, pp. 517–526. DOI:
10.1145/344779.345074.
an SVG vector image, which can be edited, and then [7] R. M. Haralick and L. G. Shapiro, Computer and robot vision. Vol. 1 Vol.
transformed into our machine-readable file, which converts 1. Place of publication not identified: Addison-Wesley, 1992.
directly to Gcode. [8] N. Adabala, N. Magnenat-Thalmann, and G. Fei, “Real-time rendering of
woven clothes,” in Proceedings of the ACM symposium on Virtual reality
software and technology, New York, NY, USA, Oct. 2003, pp. 41–47.
DOI: 10.1145/1008653.1008663.
[9] P. Khungurn, D. Schroeder, S. Zhao, K. Bala, and S. Marschner,
“Matching Real Fabrics with Micro-Appearance Models,” ACM Trans.
Graph., vol. 35, no. 1, p. 1:1-1:26, Dec. 2016, DOI: 10.1145/2818648.
[10] D. C. Banks, “Illumination in diverse codimensions,” in Proceedings of
the 21st annual conference on Computer graphics and interactive
techniques, New York, NY, USA, Jul. 1994, pp. 327–334. DOI:
10.1145/192161.192246.
[11] J. Canny, “A Computational Approach to Edge Detection,” IEEE Trans.
Pattern Anal. Mach. Intell., vol. PAMI-8, no. 6, pp. 679–698, Nov. 1986,
DOI: 10.1109/TPAMI.1986.4767851.
[12] K. J. Dana, B. V. Ginneken, S. K. Nayar, and J. J. Koenderink,
FIGURE 7: Raster Image “Reflectance and Texture of Real-World Surfaces,” ACM Trans. Graph.,
vol. 18, no. 1, p. 34.
[13] O. Deussen, S. Hiller, C. Van Overveld, and T. Strothotte, “Floating
VI. CONCLUSION Points: A Method for Computing Stipple Drawings,” Comput. Graph.
Computerized embroidery machines not only increase the Forum, vol. 19, no. 3, pp. 41–50, Sep. 2000, DOI: 10.1111/1467-
efficiency and quality of embroidered production while cutting 8659.00396.
labour and production costs, but they also broaden the varieties [14] G. Elber and G. Wolberg, “Rendering traditional mosaics,” Vis. Comput.,
2003, DOI: 10.1007/s00371-002-0175-x.
of needlework accessible, advancing human civilization and [15] W. F. Engel and Charles River Media (Firm), ShaderX7: advanced
maintaining Pakistan's traditional embroidery. We've discussed rendering techniques. Australia: Course Technology/Cengage Learning,
approaches for producing high-quality results in real time. 2009.
Strong animation tools for embroidery is, a typical textile [16] A. Hausner, “Simulating decorative mosaics,” in Proceedings of the 28th
annual conference on Computer graphics and interactive techniques,
embellishment and being able to produce and draw it New York, NY, USA, Aug. 2001, pp. 573–580. DOI:
automatically is useful. This study demonstrated the best 10.1145/383259.383327.
conversion tool for converting any embroidery file into a [17] Shaikh, Muhammad Kashif, and Palaniappan, Sellappan and Khodadadi,
machine-readable format. This research's main purpose is to Touraj, An AI-driven automotive smart black box for accident and theft
prevention, International Journal of Modelling, Identification and
turn an old sewing machine into an embroidery machine using a Control, volume 39, page 332-339, 2021.
user-interactive software application that accepts various input [18] J. Y. Han and K. Perlin, “Measuring bidirectional texture reflectance with
files and converts them to a computerized embroidery file. a kaleidoscope,” ACM Trans. Graph., vol. 22, no. 3, pp. 741–748, Jul.
2003, DOI: 10.1145/882262.882341.
[19] H. Li and D. Mould, “Structure-preserving stippling by priority-based
FUNDING STATEMENT error diffusion,” in Proceedings of Graphics Interface 2011, Waterloo,
CAN, May 2011, pp. 127–134.
73
[20] G. Winkenbach and D. H. Salesin, “Computer-generated pen-and-ink
illustration,” in Proceedings of the 21st annual conference on Computer
graphics and interactive techniques, New York, NY, USA, Jul. 1994, pp.
91–100. DOI: 10.1145/192161.192184.
[21] J. M. Kaldor, D. L. James, and S. Marschner, “Simulating knitted cloth at
the yarn level,” in ACM SIGGRAPH 2008 papers, New York, NY, USA,
Aug. 2008, pp. 1–9. DOI: 10.1145/1399504.1360664.
[22] Iqbal Hussain, Muhammad Ather, Babar Khan, Zhijie Wang, and Shenyi
Ding. 2020. "Woven Fabric Pattern Recognition and Classification Based
on Deep Convolutional Neural Networks" Electronics 9, no. 6: 1048.
https://DOI.org/10.3390/electronics9061048
[23] Kumar, K.S.; Bai, M.R. Deploying multi layer extraction and complex
pattern in fabric pattern identification. Multimed. Tools Appl. 2020, 79,
10427–10443. [Google Scholar] [CrossRef]
[24] T. Ijiri, R. Mech, T. Igarashi, and G. Miller, “An Example-based
Procedural System for Element Arrangement,” 2008, Accessed: Mar. 24,
2022. [Online]. Available:
https://diglib.eg.org:443/xmlui/handle/10.2312/CGF.v27i2pp429-436

74