Kurdistan Regional Government - Iraq

Ministry of Higher Education and Scientific Research

Salahaddin University-Erbil
College of Engineering
Directorate of Postgraduate and Scientific Affairs

Frequency Domain Transformations in

Digital Image Processing
Image Processing

Student Name
Farhang Bradosti

Higher Diploma
Salahaddin University- Erbil
College of Engineering
Software Department

12/2023
Abstract
Digital image processing plays a pivotal role in analysing and manipulating
images for various applications. One significant aspect of this field is the
utilization of frequency domain transformations, particularly the Fourier
Transform. This report aims to elucidate the fundamentals, applications, and
implications of frequency domain transformations in the realm of digital
image processing.
The introduction provides a contextual overview, distinguishing between
spatial and frequency domains. We delve into the Fourier Transform, a
mathematical tool that enables the conversion of images between these
domains. Understanding this transformation is crucial for unlocking the
hidden frequency information embedded in images.

The report then explores the application of the Fourier Transform in image
processing. It details the mathematical underpinnings of the transform and
illustrates how it is employed to analyse and manipulate images. Filtering in
the frequency domain is elucidated, showcasing the powerful impact of filters
on image characteristics. Moving forward, the focus shifts to image
enhancement techniques, where frequency domain transformations prove
invaluable. Techniques like histogram equalization and contrast stretching are
explained with practical examples, highlighting their role in improving image
quality.

Real-world applications form a pivotal part of the discussion, demonstrating

the ubiquity of frequency domain transformations in diverse fields, including
medical imaging and satellite imagery. These applications underscore the
transformative potential of these techniques in addressing real-world
challenges.

i
However, challenges are acknowledged, providing insight into the
complexities associated with implementing frequency domain
transformations. The report concludes with a look towards the future,
exploring potential advancements and trends in this dynamic field.
In essence, this report serves as a comprehensive guide to understanding and
appreciating the role of frequency domain transformations in digital image
processing. It is our hope that this exploration will equip readers with the
knowledge needed to harness the power of these transformations for
innovative applications and future developments in image processing.

ii
Contents
Introduction......................................................................................................1

Basics of Digital Image Processing: Unveiling the Visual Magic...................3

Frequency Domain: Unveiling the Harmonies in Images................................4

Fourier Transform in Image Processing: Unveiling Visual Insights................6

Filtering in Frequency Domain: Shaping Visual Realities...............................9

Image Enhancement Techniques: Transforming Pixels into Brilliance.........11

Applications: Unveiling the Real-World Magic of Frequency Domain

Transformations..............................................................................................14

Conclusion......................................................................................................16

Reference........................................................................................................17

iii
Introduction
Digital image processing, a field at the crossroads of computer science and
signal processing, has become an indispensable tool in our visually-driven
world. As we navigate through an era inundated with images, understanding
how we can process and enhance them is crucial. One fundamental aspect
that propels this field forward is the concept of frequency domain
transformations, specifically, the Fourier Transform.
At its core, digital image processing involves manipulating images using
computer algorithms. Whether it's refining medical scans, enhancing satellite
imagery, or simply applying filters on social media, digital image processing
is omnipresent. In this landscape, the concept of 'domains' is pivotal. Think of
a 'domain' as a lens through which we view and understand images.

Our journey begins with a fundamental distinction: the spatial domain, which
deals with the direct pixel values of an image, and the frequency domain, a
realm where images are expressed in terms of their frequency components.
The Fourier Transform acts as a bridge, enabling us to seamlessly traverse
between these domains. In simpler terms, it unveils hidden patterns and
frequencies within an image that are not immediately apparent.
Why does this matter? Consider the filters you use on your favourite photo-
editing app or the sharpness adjustments made to medical scans. These
manipulations often find their roots in the frequency domain. This report aims
to demystify these transformations, exploring not only their mathematical
foundations but also their practical applications.

As we embark on this exploration, we will unravel the Fourier Transform's

role in analysing images, delve into the world of frequency domain filters,
and illuminate how these transformations are harnessed for image
enhancement.

1
Real-world examples will illustrate their impact in fields ranging from
healthcare to remote sensing.
But, as with any technology, challenges abound. Navigating these challenges
is essential to fully realize the potential of frequency domain transformations.
As we conclude, we'll gaze into the future, contemplating the evolving
landscape of digital image processing and the role these transformations
might play in shaping it.

In essence, this report is an invitation to understand the magic behind the

images we encounter daily. By the end, we hope you'll see digital image
processing not as a complex labyrinth, but as a fascinating journey where
mathematics meets creativity, unlocking new dimensions in the visual realm.

2
1- Basics of Digital Image Processing: Unveiling the Visual Magic
In the ever-evolving landscape of digital advancements, we find ourselves
surrounded by a myriad of images, each telling a story or capturing a
moment. Digital Image Processing, at its core, is the art and science of
manipulating these images using computer algorithms. Why does this matter?
Because it allows us to enhance, analyse, and transform the visual
information embedded in these images.
Relevance of Digital Image Processing:
 Imagine your smartphone camera capturing a moment. Digital image
processing steps in to refine and improve the quality of that image,
making it clearer, brighter, and perhaps even more vibrant. It's the
magic wand that turns a good photo into a great one.
Spatial and Frequency Domains:
 Now, let's talk about domains - the perspectives through which we
view images. The spatial domain deals with the direct pixel values of
an image. It's the tangible, visual representation that we see on our
screens. On the other hand, the frequency domain is a bit like seeing
the music behind the image – it expresses images in terms of their
frequency components.
1. Spatial Domain Techniques: Think of this as working directly with
the pixels. Techniques like blurring, sharpening, and edge detection are
common here. For example, blurring smoothens an image, while
sharpening enhances the edges, making details pop.

3
 2. Frequency Domain Techniques: Now, enter the realm of frequencies.
Here, we introduce the Fourier Transform, a mathematical wizard that
unveils hidden patterns and frequencies within an image. Filters, which
are like magic lenses, help us enhance or suppress specific frequencies.

This journey into the basics sets the stage for our exploration of how these
domains and techniques intertwine, creating a symphony of visual
transformations. Stick with us as we demystify the visual magic of digital
image processing!

2- Frequency Domain: Unveiling the Harmonies in Images

In the world of digital image processing, there's a hidden dimension that adds
depth and richness to our visual experiences – the frequency domain. Let's
break down this concept in simple terms, exploring its role, the magic of
Fourier Transform, and the dance between spatial and frequency domains.
Defining the Frequency Domain:
 Imagine your favourite song. What if you could see its melody? The
frequency domain is a bit like that. It's a lens through which we view
images not in terms of pixels but frequencies. It reveals the underlying
patterns and harmonies that contribute to the visual composition.

4
*Figure-2: Frequency domain representation. The magnitude of the
Fourier transform as a function of horizontal and vertical frequencies is
represented in logarithmic terms (deciBeis) for ease of visualization.

The Fourier Transform and its Significance:

 Now, meet the maestro behind this transformation – the Fourier
Transform. It's a mathematical wizard that takes an image and breaks it
down into its fundamental frequency components. This allows us to see
the 'visual notes' that compose the image, much like breaking down a
musical composition into individual notes.
Mathematically, the Fourier Transform is expressed as:
F(u,v)=∬f(x,y)⋅e−j2π(ux+vy)dxdy
(Don't worry if the equation looks a bit intimidating; it's just the mathematical
magic that happens behind the scenes!)

* View of a signal in the time and frequency

domain

Conversion between Spatial and Frequency Domains:

 The intriguing part is how images seamlessly move between these
domains. The Fourier Transform acts as a bridge. It takes an image in
the spatial domain, reveals its frequency components, and vice versa.
It's akin to translating between musical notes and the melody they
create. By understanding this interplay between spatial and frequency
domains, we unlock the ability to manipulate images in ways that were
once pure
5
 imagination. Stick with us as we venture further into the realm where
images transform into visual symphonies

3- Fourier Transform in Image Processing: Unveiling Visual Insights

In our journey through digital image processing, the spotlight now falls on a
captivating performer – the Fourier Transform. Let's dive into how it not only
takes center stage but also transforms the way we perceive and enhance
images. 3.1- Application in Digital Image Processing:
Machine Learning, a field that focuses on developing algorithms capable of
learning from data, has seen a growing use of the Fourier Transform. Its
applications in this domain are diverse and impactful:
1. Time Series Analysis: In fields like finance, healthcare, and weather
forecasting, time series data is abundant. The Fourier Transform can
help extract relevant features from time series data by analysing its
frequency components. This is crucial for tasks like anomaly detection,
trend analysis, and forecasting.
2. Natural Language Processing: Text data, when represented as a
sequence of words, can be treated as a discrete signal. By applying the
Fourier Transform, it becomes possible to analyse the textual data in
the frequency domain, which has applications in text classification,
sentiment analysis, and topic modelling.
3. Feature Engineering: Feature engineering is a critical step in the
machine learning pipeline. By converting data into the frequency
domain, valuable features can be extracted that might be difficult to
capture in the time domain. This can lead to more robust and accurate
machine learning models.

6
3.2-Mathematical Foundation:
Now, let's peek behind the curtain at the math behind this magic. The Fourier
Transform takes an image and breaks it down into its fundamental
frequencies.
The mathematical dance looks like this: F(u,v)=∬f(x,y)⋅e−j2π(ux+vy)dxdy
(It may seem like a complex equation, but trust us, it's the wizardry that turns
pixels into visual symphonies!)
The Fourier Transform is a mathematical operation that transforms a function
from the time or spatial domain to the frequency domain. The continuous
Fourier Transform is defined as follows:
F(u)=∫−∞∞f(t)⋅e−j2πutdt
In this equation:
 F(u) represents the complex amplitude of the frequency component at
a given frequency u in the frequency domain.
 f(t) is the function in the time or spatial domain.
 e−j2πut is the complex exponential function.
This equation essentially expresses the idea that any signal, image, or
function in the time or spatial domain can be represented as a sum of
sinusoidal functions of different frequencies in the frequency domain.
For discrete signals or images sampled at discrete points, the discrete Fourier
Transform (DFT) is often used, and it is represented as:
X(k)=∑n=0N−1x(n)⋅e−j2πNkn
Here:
 X(k) is the complex amplitude of the frequency component at a given
discrete frequency k in the frequency domain.
x(n) is the value of the signal or image at the nth sample point.
 N is the total number of samples.
The Fast Fourier Transform (FFT) is an algorithm for efficiently computing the
DFT, making the computation more practical for real-world applications.

7
3.3- Examples of Image Analysis:
 The true beauty emerges when we see how the Fourier Transform is
applied in image analysis. Consider medical imaging; the Fourier
Transform unveils crucial details in scans, aiding diagnoses. In satellite
imagery, it helps identify patterns that might be missed by the naked eye.
 Another everyday example is compression. The Fourier Transform plays
a role in compressing images, making them easier to store and transmit
without losing vital information.

* Compression performance among wavelet, Fourier and SVD.

By grasping how the Fourier Transform works, we unlock the ability to

decode the visual language embedded in images. Stick with us as we
explore how this mathematical maestro transforms mere pixels into a visual
masterpiece!

8
4- Filtering in Frequency Domain: Shaping Visual Realities
As we continue our exploration of the digital image processing realm, we
come across a powerful tool that allows us to sculpt and enhance visual
information
– filtering in the frequency domain. Let's embark on a journey to understand
the concept, the application through the Fourier Transform, and witness the
impact of common frequency domain filters.
4.1- Concept of Filtering in Frequency Domain:
Think of filtering as adjusting the visual 'ingredients' of an image. In the
frequency domain, we manipulate the various frequencies that make up an
image. Just like tuning a radio to focus on a specific wavelength, we use
filters to emphasize or suppress certain visual features.

Low pass filter: Low pass filter removes the high frequency components that
means it keeps low frequency components. It is used for smoothing the
image. It is used to smoothen the image by attenuating high frequency
components and preserving low frequency
components. Mechanism of low pass filtering in frequency
domain is given by:
G(u, v) = H(u, v) . F(u, v)
where F(u, v) is the Fourier Transform of original image
and H(u, v) is the Fourier Transform of filtering mask
High pass filter: High pass filter removes the low frequency components that
9
means it keeps high frequency components. It is used for sharpening the
image. It is used to sharpen the image by attenuating low frequency
components and

1
preserving high frequency components. Mechanism of high pass filtering in
frequency domain is given by:
H(u, v) = 1 - H'(u, v)
where H(u, v) is the Fourier Transform of high pass filtering
and H'(u, v) is the Fourier Transform of low pass filtering
Band pass filter: Band pass filter removes the very low frequency and very
high frequency components that means it keeps the moderate range band of
frequencies. Band pass filtering is used to enhance edges while reducing the
noise at the same time.
4.2- Applying Filters using the Fourier Transform:
Now, enter the Fourier Transform as our maestro. Filters act as its
instruments, and the image is our canvas. The Fourier Transform reveals the
frequency components, and filters modify these components selectively.
High-pass filters might enhance edges, while low-pass filters can

smoothen and blur.

The process involves multiplying the Fourier Transform of the image by the
frequency response of the desired filter. This selective manipulation of
frequencies is what gives us control over the visual outcome.
The mathematical operation can be expressed as: G(u,v)=H(u,v)⋅F(u,v)
where: G(u,v) is the filtered image in the frequency domain, H(u,v) is the
frequency response of the filter, and F(u,v) is the Fourier Transform of the
1
original image.

1
4.3- Examples of Frequency Domain Filters:
 Let's bring this to life with examples. The Gaussian filter, for instance,
is like applying a gentle blur. It smoothens out an image, reducing high-
frequency noise.

 On the other hand, the high-pass filter sharpens an image by

emphasizing edges and details.
 Combining these filters allows for nuanced adjustments, giving us the
power to craft images to our liking. In essence, filtering in the
frequency domain is our artistic brush, allowing us to mold visual
reality with precision. Stick with us as we continue to unravel the
canvas of digital image processing!

5- Image Enhancement Techniques: Transforming Pixels into Brilliance

In the vast canvas of digital image processing, the pursuit of refining and
elevating images to their full potential takes center stage. As we navigate this
artistic journey, we discover the transformative capabilities of frequency
domain transformations, akin to a magic wand that reshapes visual
landscapes. Join us on this exploration where we delve into techniques such
as histogram equalization and contrast stretching in the frequency domain,
witnessing first hand their profound impact on image quality.

1
Employing Frequency Domain Transformations for Enhancement:
Imagine frequency domain transformations as the sculptor's chisel and
hammer, delicately carving and reshaping visual landscapes. In the realm of
image enhancement, these transformations emerge as the architects of
change, capable of unveiling hidden details and enriching the overall
quality of an image.
Histogram Equalization:
Let's take a deeper dive into the technique of histogram equalization, a
process akin to giving each pixel a chance to shine. In the frequency domain,
this method redistributes pixel intensities, fostering a balanced and well-
distributed histogram. The result? An image with heightened contrast and
improved visibility of intricate details.
Mathematically, this process is expressed as:
G(u,v)=H(u,v)⋅F(u,v)
Where:
 G(u,v) is the enhanced image,
 H(u,v) is the frequency response, and
 F(u,v) is the Fourier Transform of the original image.
Contrast Stretching:
Consider contrast stretching as an opportunity to infuse a vibrant tune into a
previously muted melody. In the frequency domain, this technique involves
stretching the range of pixel intensities, intensifying contrast, and rendering
the image visually richer.
Mathematically, contrast stretching is expressed as a simple linear operation
on pixel intensities:
G(u,v)=a⋅F(u,v)+b
Where:
 a and b are constants that wield control over the stretching effect.

1
]

Impact on Image Quality:

Now, let's witness the tangible impact of these techniques. The following
examples showcase the original image, the Fourier Transform, and the
enhanced image after applying histogram equalization and contrast
stretching.

Observe how the details become more pronounced, and the overall quality of
the image ascends through these frequency domain enhancement techniques.
In essence, frequency domain transformations empower us to breathe life into
pixels, transcending ordinary images into extraordinary visual experiences.
Stick with us as we continue our expedition through the diverse and vibrant
palette of digital image processing!

1
6- Applications: Unveiling the Real-World Magic of Frequency Domain
Transformations
In our exploration of digital image processing, we uncover the tangible and
transformative impact of frequency domain transformations across various
real-world applications. Join us as we shine a spotlight on crucial areas such
as medical imaging, satellite imagery, and more, where these techniques play
a pivotal role.
1. Medical Imaging:
 Picture the intricate details within the human body, vital for accurate
diagnoses. Frequency domain transformations are indispensable in
medical imaging, enhancing the quality of scans and unveiling subtle
nuances that might escape the naked eye. Whether it's identifying
abnormalities in X-rays or refining MRI images, these transformations
contribute to the precision and reliability of medical diagnostics.

2. Satellite Imagery:
 Imagine looking down from space and decoding valuable information for
agriculture, urban planning, or disaster response. Frequency domain
transformations prove to be the guiding force in satellite imagery. They
help identify patterns, enhance features, and extract critical data that aids
in decision-making. From monitoring crop health to assessing
environmental changes, these techniques contribute to a clearer, more
insightful view from above.

1
3. Wireless Communication:
 Consider the seamless transfer of information through wireless channels.
Frequency domain transformations are instrumental in modulating and
demodulating signals, optimizing communication systems for efficient
data transmission. These techniques ensure that the information sent and
received is clear, reliable, and free from interference, contributing to the
robustness of wireless communication networks.
4. Speech and Audio Processing:
 Think about the clarity of sound in your favorite podcast or the precision
of voice recognition systems. Frequency domain transformations play a
vital role in speech and audio processing. From noise reduction to
enhancing specific frequencies, these techniques refine audio signals,
delivering a richer and more intelligible auditory experience.
In essence, frequency domain transformations are the unsung heroes behind
the scenes, shaping the clarity and reliability of information in diverse fields.
From the intricacies of the human body to the vast expanse of space, these
techniques enhance our ability to interpret, analyse, and make informed
decisions in the complex landscape of digital image processing. Stick with us
as we continue to unveil the multifaceted applications of these transformative
tools!

1
Conclusion:

As we draw the curtains on our exploration of frequency domain

transformations in digital image processing, we find ourselves in awe of the
symphony that unfolds when pixels meet mathematics. This journey has
taken us from the basics of reshaping visual landscapes to the profound
applications that echo in real-world scenarios. Let's distil our insights into a
harmonious conclusion.
At the heart of this exploration lies the transformative power of frequency
domain transformations. Like an artist with a palette of colours, these
techniques allow us to breathe life into pixels, revealing hidden details and
enhancing the visual allure of images. Behind the scenes, mathematical
expressions and functions dance to orchestrate the magic. The Fourier
Transform, our maestro, takes the lead, breaking down images into their
fundamental frequencies and enabling a seamless transition between spatial
and frequency domains. F(u,v)=∬f(x,y)⋅e−j2π(ux+vy)dxdy
We've witnessed the elegance of histogram equalization and contrast
stretching, where frequency domain transformations become the architects of
enhanced image quality. Each technique, a brushstroke on the canvas of
visual perception.
The applications of frequency domain transformations extend far beyond the
digital realm. From the precision of medical diagnoses to decoding patterns in
satellite imagery, these techniques resonate in fields such as wireless
communication, audio processing, and finance. As we conclude this
exploration, we look towards a future where advancements in digital image
processing will continue to be shaped by the harmonies of frequency domain
transformations. Challenges will be met with innovation, and the canvas of
visual expression will evolve.

1
In essence, frequency domain transformations stand as a testament to the
dynamic interplay between mathematics and artistry, unveiling a world where
pixels transcend their static existence. As we bid adieu to this symphony, let
the echoes of frequency domain transformations linger, reminding us that in
the world of digital image processing, every pixel has a story, waiting to be
told.

Reference:
 Gonzalez, R. C., & Woods, R. E. (2008). "Digital Image Processing."
Pearson.
 Pratt, W. K. (2007). "Digital Image Processing: PIKS Inside." John Wiley
& Sons.
 Oppenheim, A. V., & Schafer, R. W. (1968). "Digital Signal Processing."
Proceedings of the IEEE, 56(5), 519-541.
 Bracewell, R. N. (1999). "The Fourier Transform and Its Applications."
McGraw-Hill.
 "Digital Image Processing Using MATLAB" by Gonzalez, Woods, and
Eddins:
 "Image Processing, Analysis, and Machine Vision" by Milan Sonka,
Vaclav Hlavac, and Roger Boyle: