ADVANCED SKIN CATEGORY

PREDICTION SYSTEM FOR

COSMETIC SUGGESTION USING
DEEP CONVOLUTIONAL NEURAL
NETWORK

PROJECT REPORT

Submitted By

ABISHEK R 710719104004
AJEYAN T.S.R 710719104009
ARAVINTH N 710719104014
GOKUL M 710719104034

In partial fulfillment for the award of the degree

of

BACHELOR OF ENGINEERING
in
COMPUTER SCIENCE AND ENGINEERING

Dr. N.G.P. INSTITUTE OF TECHNOLOGY

(An Autonomous Institution)
(Approved by AICTE, New Delhi & Affiliated to Anna University)
Recognized by UGC, Accredited by NAAC with A+ & NBA (BME, CSE, ECE, EEE & MECH)
Coimbatore – 641 048.

ANNA UNIVERSITY: CHENNAI - 600 025

APIRIL 2023
 ANNA UNIVERSITY: CHENNAI - 600 025

BONAFIDE CERTIFICATE

Certified that this Project report “ADVANCED SKIN

CATEGORY PREDICTION SYSTEM FOR COSMETIC SUGGESTION
USING DEEP CONVOLUTIONAL NEURAL NETWORK” is the bonafide
work of ABISHEK R - 710719104004, AJEYAN T.S.R – 710719104009,
ARAVINTH N - 710719104014, GOKUL M - 710719104034 who carried out
the project work under my supervision.

HEAD OF THE DEPARTMENT SUPERVISOR

Dr. D. PALANIKKUMAR M.E., Ph.D. Dr. B. DHIYANESH, M.Tech, Ph.D.,
Professor and Head, Associate Professor,
Department of Computer Science and Department of Computer Science and
Engineering, Engineering,
Dr. N. G. P. Institute of Technology, Dr. N. G. P. Institute of Technology,
Coimbatore - 641 048. Coimbatore - 641 048.

Submitted for the Project viva-voce examination held on __________

INTERNAL EXAMINER EXTERNAL EXAMINER

 TABLE OF CONTENTS

CHAPTER NO TITLE PAGE NO

ABSTRACT – ENGLISH i

ABSTRACT - TAMIL ii

LIST OF FIGURES iii

ABBREVIATIONS iv

1 INTRODUCTION 1

1.1 INTRODUCTION 1
COMPOSITION OF COSMETIC 1
1.2
PRODUCTS
1.3 MACHINE LEARNING 1

1.4 EVOLUTION OF MACHINES 2

1.5 HOW MACHINE LEARNING 2

WORKS
1.6 REQUIREMENTS OF DEEP 3
LEARNING
1.7 VARIOUS DEEP LEARNING 3
TECHNIQUES

2 LITERATURE SURVEY 5

3 PROPOSED METHODOLOGY 9

3.1 PROPOSED WORK 9

3.2 DATASETS 10
3.2.1 Image labelling and Dataset 10
Distributions
 3.3 PRE-PROCESSING 10
3.3.1 Pre-processing Steps 11
3.4 FEATURE EXTRACTION 11
3.5 ALGORITHM USED CNN 11

4 SYSTEM SPECIFICATION 14
4.1 HARDWARE REQUIREMENTS 14
4.2 SOFTWARE REQUIREMENTS 14
4.3 TECHNOLOGY AND TOOLS USED 14

5 TECHNOLOGIES AND TOOLS 15

DESCRIPTION
5.1 SOFTWARE DESCRIPTION 15
5.1.1 Python 15
5.2 DEEP LEARNING 17
5.3 LIBRARIES DESCRIPTION 18
5.3.1 Activation Function 18
5.3.2 NumPy 18
5.3.3 TensorFlow 18
5.3.4 TensorFlow Implementation 19
5.3.5 Data Parallel Training 20
5.3.6 Keras 21
5.3.7 Matplotlib 22
5.3.8 Cv2 23
5.4 DATA TYPE OBJECTS 23
5.5 TOOLS DESCRIPTION 23
5.5.1 Thonny IDE 23
6 PROPOSED MODULES 25
6.1 COLLECTION OF DATASETS 25
6.2 DATASETS PRE-PROCESSING 25
6.3 TRAINING THE MODEL 25
6.4 TESTING PHASE 26

7 IMPLEMENTATION 27
7.1 PRE-PROCESSING AND FEATURE 27
EXTRACTION
7.2 TRAINING AND VALIDATION 27
7.3 TESTING AND COSMETIC 27
SUGGESTION

8 TESTING 29
8.1 TESTING INTRODUCTION 29
8.2 TYPES OF TESTING CONSIDERED 29

9 RESULT AND DISCUSSION 32

10 CONCLUSION AND FUTURE WORK 34

10.1 CONCLUSION 34

10.2 FUTURE SCOPE 34

REFERENCE 35
APPENDICIES – SCREEN SHOT 38

APPENDICIES – SOURCE CODE 44

 DECLARATION

I hereby declare that the project work entitled “ADVANCED SKIN

CATEGORY PREDICTION SYSTEM FOR COSMETIC SUGGESTION
USING DEEP CONVOLUTIONAL NEURAL NETWORK”, submitted to
the autonomous project Viva Voce – April 2023 in partial fulfillment for the
award of the degree of “BACHELOR OF ENGINEERING IN COMPUTER
SCIENCE AND ENGINEERING”, is the report of the original project work
done by me under the guidance of Dr. B. DHIYANESH, M. Tech, Ph. D.,
Associate Professor, Department of Computer Science and Engineering,
Dr. N.G.P. Institute of Technology, Coimbatore - 641 048.

NAME SIGNATURE
ABISHEK R
AJEYAN T.S.R

ARAVINTH N
GOKUL M

I certify that the declaration made by the above candidate is true.

Project Guide
Dr. B. DHIYANESH, M. Tech, Ph. D.,
Associate Professor,
Department of Computer Science and Engineering,
Dr. N.G.P. Institute of Technology,
Coimbatore - 641 048.
 ACKNOWLEDGEMENT

Words act as gateway to express tokens of acknowledgement. First of all,

we would like to thank the supreme power, the Almighty God, who has given us
the strength and courage to complete our work successfully.
We express our profound gratitude and deep sense of thanks
to Dr. Nalla G. Palaniswami, MD., AB (USA), Chairman of Kovai Medical
Center & Hospital, for providing us with the necessary facilities to complete our
project work effectively.
Our heartfelt gratitude to Dr. Thavamani D Palaniswami, MD., AB
(USA), Secretary of Dr. N. G. P. Institute of Technology, for her generous
attitude and constant motivation, which had been one of the sole reasons to
complete our project.
We are sincerely grateful to Dr. S. U. Prabha, M.E., Ph.D., Principal,
who has always been a source of inspiration, well-wisher and a pillar of support
for all the students in our institution by rendering full motivation whenever
required.
We are highly indebted to Dr. D. Palanikkumar, M.E., Ph.D., Head of
Department, Department of Computer Science and Engineering, for his
dedication, keen interest, meticulous scrutiny and overwhelming attitude to help
her students that have helped us to a very great extent to accomplish this task.
We wish to thank our Project Coordinator Dr. B. Dhiyanesh, M.Tech.,
Ph.D., Associate Professor, Department of Computer Science and
Engineering, for his excellent assistance, aspiring guidance, regular feedback
and invaluably constructive ideas.
Our sincere gratitude to our Supervisor Dr. B. Dhiyanesh, M.Tech.,
Ph.D., Associate Professor, Department of Computer Science and
Engineering for willingly share his precious time by giving us useful comments,
remarks, timely suggestion with kindness, enthusiasm and dynamism have
enabled us to complete our thesis. Finally, we owe huge thanks to our Parents, all
the Faculty members and our classmates, whose love and insights have so deeply
enriched our work.
 ABSTRACT
ENGLISH

Nowadays, cosmetic products are quite important to a person's look. Customers

get access to a variety of products through online shopping and e-commerce
websites. We find it challenging to choose the ideal product for our skin. As such,
we propose a system that will allow us to accurately determine what product is
suitable for our skin type. The composition depends on skin types, which might
be dry, oily, or normal. The composition proposal will be superior to the
established method. We will make the difficult task for the IT sector in cosmetic
and beauty care easier by utilizing deep learning techniques. The beauty sector is
expanding rapidly over time, which has a positive knock-on effect on both the
products it offers and its customers. The rapid growth of products and customers
underscores the importance of choosing the right cosmetic product. There is a
necessity to choose the ideal cosmetic product for oneself based on personal
factors because cosmetics are important to one's personality (i.e., skin type). As
every person has a different skin texture, it might be challenging to find the right
cosmetic for a customer's skin type. Even if we locate the right product for the
consumer, skin problems can still arise because of very complex issues. For
solving this problem in the beauty industry, we can use AI algorithms because
they provide better results when working with vast amounts of unstructured data.

i
 ABSTRACT
TAMIL

இன் று, அழகுசாதனப் பபாருட்கள் ஒரு நபரின் ததாற் றத்திற் கு மிகவும்

முக்கியம் . ஆன்லைன் ஷாப் பிங் மற் றும் இ-காமர்ஸ் இலையதளங் கள் மூைம்
வாடிக்லகயாளர்கள் பை் தவறு தயாரிப் புகளுக்கான அணுகலைப்
பபறுகின்றனர். நமது சருமத்திற் கு ஏற் ற பபாருலளத் ததர்ந்பதடுப்பது
சவாைாக இருக்கிறது. எனதவ, எங் கள் ததாை் வலகக்கு எந்த தயாரிப் பு
பபாருத்தமானது என்பலத துை் லியமாக தீர்மானிக்க அனுமதிக்கும் ஒரு
அலமப் லப நாங் கள் முன் பமாழிகிதறாம் . கைலவயானது ததாை்
வலககலளப் பபாறுத்தது, அலவ உைர்ந்த, எை்பைய் கூட்டு அை் ைது
இயை் பானதாக இருக்கைாம் . ஆழ் ந்த கற் றை் நுட்பங் கலளப்
பயன்படுத்துவதன் மூைம் , ஒப் பலன மற் றும் அழகுப் பராமரிப் பிை் ஐடி
துலறயின் கடினமான பைிலய எளிதாக்குதவாம் . அழகுத் துலறயானது
காைப் தபாக்கிை் தவகமாக விரிவலடந்து வருகிறது, இது வழங் கும்
தயாரிப் புகள் மற் றும் அதன் வாடிக்லகயாளர்கள் ஆகிய இரை்டிலும்
சாதகமான தாக்கத்லத ஏற் படுத்துகிறது. தயாரிப்புகள் மற் றும்
வாடிக்லகயாளர்களின் விலரவான வளர்ச்சி சரியான ஒப் பலன
தயாரிப் லபத் ததர்ந்பதடுப் பதன் முக்கியத்துவத்லத அடிக்தகாடிட்டுக்
காட்டுகிறது. ஒருவரின் ஆளுலமக்கு (அதாவது ததாை் வலக) அழகுசாதனப்
பபாருட்கள் முக்கியமானலவ என்பதாை் தனிப் பட்ட காரைிகளின்
அடிப் பலடயிை் தனக்கான சிறந்த ஒப்பலனப் பபாருலளத் ததர்ந்பதடுக்க
தவை்டிய அவசியம் உள் ளது. ஒவ் பவாரு நபருக்கும் பவவ் தவறு ததாை்
அலமப் பு இருப் பதாை் , வாடிக்லகயாளரின் ததாை் வலகக்கு சரியான
அழகுசாதனத்லத கை்டுபிடிப் பது சவாைாக இருக்கைாம் . நுகர்தவாருக்கு
சரியான தயாரிப்லப நாம் கை்டறிந்தாலும் , மிகவும் சிக்கைான
பிரச்சலனகளாை் ததாை் பிரச்சலனகள் இன் னும் எழைாம் . அழகுத் துலறயிை்
இந்தச் சிக்கலைத் தீர்க்க, நாம் AI அை் காரிதம் கலளப் பயன்படுத்தைாம் ,
ஏபனனிை் அலவ பபரிய அளவிைான கட்டலமக்கப் படாத தரவுகளுடன்
பைிபுரியும் தபாது சிறந்த முடிவுகலள வழங் குகின்றன.

ii
 LIST OF FIGURES

FIGURE NO. FIGURE NAME PAGE NO.

3.1 Block Diagram 9

3.2 Flow Diagram 13
9.1 Training and Validation Loss 33
9.2 Training and Validation Accuracy 33
9.3 Confusion Matrix 34
A1 Classification of Datasets 38
A2 Pre-Processing and Feature Extraction 38
A3 Importing Libraries 39
A4 Training and Accuracy 39
A5 Cosmetics Products Used in Application 40
A6 User Interface 40
A7 Suggestion of Cosmetics for Dry Skin 41
A8 Suggestion of Cosmetics for Oily Skin 41
A9 Suggestion of Cosmetics for Cosmetics 42
Skin
A10 Suggestion of Cosmetics for Normal Skin 42

iii
 LIST OF ABBREVIATIONS

AI - Artificial Intelligence
AE - Auto Encoder
ANN - Artificial Neural Networks
API - Application Programming Interface
CNN - Convolutional Neural Network
CPU - Central Processing Unit
CSV - Comma-Separated Values
DBN - Deep Belief Network
DNN - Deep neural networks
DOG - Difference of Gaussian
DTM - Document Term Matrix
GPU - Graphics Processing Unit
ML - Machine Learning
PACS - Picture Archiving Communication System
RAM - Random Access Memory
RELU - Rectified Linear Unit
RGB - Red Green Blue
RNN - Recurrent Neural Networks
SDA - Stacked Denoising Auto Encoders
SVM - Support Vector Machines
UI - User Interface
VAE - Variational Autoencoders
VGG - Visual Geometry Group

iv