THE IMAGE COLORIZATION

TECHNIQUES USING DEEP LEARNING

A PROJECT REPORT

Submitted by

ABIYEASER A

(210522243002)
in partial fulfillment of the Requirements for the award of the degree

BACHELOR OF TECHNOLOGY
IN

ARTIFICIAL INTELLIGENCE AND DATA

SCIENCE

DMI College of Engineering,

Palanchur, Chennai

DECEMBER 2024

i
 DECLARATION

I undersigned hereby declare that the project report "IMAGE COLORIZATION

USING DEEP LEARNING" submitted for partial fulfillment of the requirements for
the award of the degree of Bachelor of Technology in Artificial Intelligence and Data
Science at DMI College of Engineering, is a Bonafide work done by us under the
supervision of Mrs. Anjali Devi A J. This submission represents my ideas in my own
words, and where ideas or words of others have been included, I have adequately and
accurately cited and referenced the original sources.

I also declare that I have adhered to the ethics of academic honesty and
integrity and have not misrepresented or fabricated any data, idea, fact, or
source in my submission. I understand that any violation of the above will be a
cause for disciplinary action by the institute and/or the University and can also
evoke penal action from the sources which have thus not been properly cited or
from whom proper permission has not been obtained. This report has not been
previously formed the basis for the award of any degree, diploma, or similar
title of any other University.

Chennai
ABIYEASER A
December
2024

ii
 DEPARTMENT OF ARTIFICIAL INTELLIGENCE
AND DATA SCIENCE
DMI COLLEGE OF ENGINEERING, PALANCHUR

CERTIFICATE

This is to certify that report entitled IMAGE COLORIZATION

USING LEARNING Submitted by ABIYEASER (210522243002) to the
DMI College of Engineering in partial fulfilments of the requirements
for the award of the degree of Bachelor of Technology in (Artificial
Intelligence and Data Science ) is a Bonafide record of the project word
carried out by her under our guidance and supervision .this report in
any form has not been submitted to any other University or Institute for
any purpose

Mrs. Anjali Devi A J Dr. K Kalai Kumar

Internal Supervisor Head of the Department
Dept. of Artificial Intelligence Dept. of Artificial Intelligence &
& DataScience DataScience

iii
 CONTENT

Contents page no

ACKNOWLEDGMENT........................................................................................5
LIST OF TABLES................................................................................................6
Chapter 1: INTRODUCTION
1.1 Introduction. ................................................................................. 7 1.2
Abstract.........................................................................................8 Chapter 2:
LITERATURE SURVEY ............................................................ 9 Chapter 3:
METHODOLOGY.............................................................................10
3.1 Process.............................................................................................11
3.2 Code.................................................................................................12
Chapter 4: RESULT...............................................................................................14
4.1 Screenshot..........................................................................................15
4.2 result conclusion................................................................................16
Chapter 5: CONCLUSION...................................................................................17

iv
 ACKNOWLEDGEMENT

First and foremost, I express my deepest thanks to our principal, Dr.

Sujatha Jamuna Anand, for providing us with all facilities for the completion of
the work. I am extremely happy to mention a great word of gratitude to Dr. K.
Kalai Kumar, Head of the Department, Department of Artificial Intelligence and

Data Science, for his aspiring guidance, invaluably constructive criticism and
friendly advice during the work.

I am deeply indebted to my project guide, Mrs. Anjali Devi A J, Assistant Professor,

Department of Artificial Intelligence and Data Science, for the support and
encouragement provided by her to make this a success.

Finally, yet importantly, I would also extend my gratefulness to all the staff
members in the Department. I also thank all my friends and well-wishers who
greatly helped me in this endeavor.

v
 LIST OF TABLES

No. Title Page No.

3.1 Literature survey in tabular form 3

vi
 CHAPTER 1
INTRODUCTION

1.1 INTRODUCTION

What is Image Colorization: Image colorization is the process of adding color t

grayscale or black-and-white images. It transforms monochromatic images into
color images, effectively restoring or enhancing their visual appeal and usefulness.
Th process can be done manually or through automated methods, such as using
machin learning

Image colorization holds immense potential not only for restoring and enhancing
historical photos but also for practical applications such as media production, design,
even medical imaging. The process provides a new way of interacting with the past,
offering a richer and more relatable visual experience.

While automatic image colorization has made great strides, challenges remain,
particularly in terms of achieving perfect color accuracy and handling complex scenes.

Nonetheless, with the continued advancement of deep learning and computational

power, the future of image colorization is poised to open up exciting possibilities in a
variety of fields.

1
 1.2 ABSTRACT

Image colorization is the process of adding color to grayscale or black-and-white

images, transforming them into realistic, full-color representations. This process
has gained significant attention in recent years due to advances in deep learning
and neural networks, particularly Generative Adversarial Networks (GANs),

which have improved the quality and accuracy of automatic colorization.

Traditionally, colorization required manual effort by artists, but modern
techniques now allow for automated solutions that can produce high-quality
results with minimal human intervention. Image colorization has applications in
various fields, including historical image restoration, film and media industries,
and artistic expression. By reviving old photographs, paintings, and motion
pictures, colorization offers a deeper understanding of visual history and context,
making these works more relatable and emotionally engaging to contemporary
audiences.

Furthermore, it holds potential for practical applications such as enhancing

medical imaging and satellite images. Despite the progress made in automated
colorization, challenges remain in achieving consistent and true-to-life color
reproduction, especially in complex or ambiguous scenes. The success of
colorization largely depends on the quality of the training data and the
sophistication of the model used. In this context, the development of more
advanced algorithms and refined techniques continues to enhance the realism and
accuracy of colorized images.

2
 Chapter 2

Literature Survey
 CHAPTER 3
METHODOLOGY
3.1 PROCESS:

Data Collection and Preprocessing

Data Collection: The first step involves gathering a large dataset of color
images paired with their grayscale counterparts.

2. Model Selection
Tra•ditional Methods: Early approaches to image coloriz ation relied on hand-crafted
features and algorithms.

Split the Data into Training and Testing Sets

Linear Regression Model (Simple Linear Regression or Multiple Linear Regression) n

deep learning, linear regression can be implemented using a simple neural network
with a single linear layer (no activation).

Train the Model

ColorizationProcess

5. Evaluation and Post-Processing

7. Challenges and Future Directions

• Accuracy of Colors: One of the main c•h allenges in image colorization is

ensuring that the predicted colors match the original objects or scene's real-
world colors. The model may misinterpret complex scenes with ambiguous
color relationships.
5
3.2 CODE:

!pip install torch torchvision fastai

!pip install deoldify

from deoldify import DeOldify from deoldify import device

import torch from fastai.vision import * import matplotlib.pyplot
as plt

if
torch.cuda.is_available():
device =
torch.device('cuda')
else:

device =
torch.device('cpu')

print(f"Device set to:

{device}")

# Step 4: Load DeOldify

model
colorizer =
DeOldify(device=device)

input_image_path='path_
5
to_your_grayscale_image. jpg'
input_image =
open_image(input_image
_path)

colorized_image =
colorizer.get_image_colori
zation(input_image)

plt. imsho w( color ized_ i m a

ge)
plt.axis('off') plt.show() colorized_image.save('col
orized_image.jpg')

5
 CHAPTER 4

RESULT

SCREENSHOT:
 Chapter 5

CONCLUSION

The conclusion about image colorization is that it is a transformative process that

brings new life to black-and-white or grayscale images, making them more accessible,
engaging, and relatable to modern audiences. Whether done manually or through
automated techniques such as deep learning, colorization has the potential to
enhance the emotional and historical impact of images, offering viewers a more vivid
connection to the past. While automated methods, particularly those using machine
learning, have made significant progress in replicating realistic colors, they still face
challenges in achieving the same level of accuracy and detail that manual colorization
or expert interpretation can provide.

Image colorization can serve as both an educational tool and a form of

artistic expression, but it requires careful attention to accuracy and context to
ensure it remains true to the original subject and time period. Image colorization is a
powerful tool that enhances grayscale images, making them more engaging and
relatable. While automated techniques like machine learning have made significant
strides, challenges remain in achieving perfect accuracy.

Overall, colorization enriches historical and artistic content, offering both emotional

impact and educational value, but requires careful attention to detail and context for
the best results.