Gender and Age Detection

Department of Computer Science and Engineering

Bangladesh University of Business and Technology

Dhaka-1216

November 2021
 Gender and Age Detection

A report
submitted to the department of Computer Science and Engineering
in partial fulfillment of the requirements
for the degree of
Bachelor of Science in Computer Science and Engineering

By

Md.Mahedy Hasan (ID:18191103048)

Mahmudul Hasan (ID:18191103057)
Mahady hasan (ID:18191103061)

Momin Ali (ID: 38-17182103011)

Mohammad Tanvir Hossain Reza (ID:38-17182103020)

Supervised by
MD. Shahiduzzaman
Assistant Professor

Department of Computer Science and Engineering (CSE)

Bangladesh University of Business and Technology (BUBT)
Mirpur-2, Dhaka-1216, Bangladesh
 ABSTRACT

Nowadays research has explored extracting auxiliary information from various biometric
techniques such as fingerprints, face, iris, palm, voice etc. This information contains some features
like gender, age, beard, moustache, scars, height, hair, skin color, glasses, weight, facial marks,
tattoos etc. All this information contributes more and more during identification. The major
changes that come across face recognition is to find the age & gender of the person. This paper
contributes a significant survey of various face recognition techniques for finding the age and
gender. The existing techniques are discussed based on their performances. This paper also
provides future directions for further research. Age and gender, two of the key facial attributes,
play a very foundational role in social interactions, making age and gender estimation from a single
face image an important task in intelligent applications, such as access control, human-computer
interaction, law enforcement, marketing intelligence and visual surveillance.
This project is based upon computer vision and the various terminologies used to process images
and detect age and gender of the person from the image.
A Convolutional Neural work is a deep neural network (DNN) widely used for the purposes of
image recognition and processing and NLP. Also known as a ConvNet, a CNN has input and
output layers, and multiple hidden layers, many of which are convolutional. In a way, CNNs are
regularized multilayer perceptrons. Computer Vision is the field of study that enables computers
to see and identify digital images and videos as a human would. The challenges it faces largely
follow from the limited understanding of biological vision. A fast and efficient gender and age
estimation system based on facial images is developed. There are many methods that have been
proposed in the literature for the age estimation and gender classification. However, all of them
have still disadvantage such as not complete reflection about face structure, face texture. This
technique applies to both face alignment and recognition and significantly improves three aspects
Within a given database, all weight vectors of the persons within the same age group are averaged
together. A range of an age estimation result is 15 to 70 years old, and divided into 13 classes with
5 years old range. Experimental results show that better gender classification and age estimation.
age and gender classification has become applicable to an extending measure of applications,
particularly resulting to the ascent of social platforms and social media. Regardless, execution of
existing strategies on real-world images is still fundamentally missing, especially when
considering the immense bounce in execution starting late reported for the related task of face
acknowledgment. In this paper we exhibit that by learning representations through the use of
significant Convolutional Neural Systems (CNN), a huge augmentation in execution can be
acquired on these errands. To this end, we propose a direct Convolutional Neural System
engineering can be used despite when the measure of learning data is limited. We survey our
procedure on the recent Adience benchmark for age and gender estimation and demonstrate it to
radically outflank current state of-the-art methods.
 DECLARATION

We declare that this thesis and the work presented in it are our own and has been generated by us
as the result of our own original research

We confirm that:

● This Work is done wholly or mainly while in candidature for a research degree at this
University.
● This report work has not been previously submitted for any degree at this university or
any other educational institutes.
● We have quoted from the work of others; the source is always given. With the exception
of such quotations, this thesis is entirely our own work.

---------------------- ---------------------- ----------------------

Md. Mahedy Hasan Mahmudul Hasan Mahady Hasan
ID: 18191103048 ID: 18191103057 ID: 18191103061
Intake: 40 Intake: 40 Intake: 40

---------------------- ----------------------
Mohammad Tanvir Hossain Reza Momin Ali
ID: 17182103020 ID: 171812103011
Intake: 38 Intake: 38
 CERTIFICATE

This is to certify that Md. Mahedy Hasan, Mahmudul Hasan, Mahady Hasan, Mohammad Tanvir
Hossain Reza and Momin Ali students of B.Sc. in CSE have completed their report work titled
“Age and Gender Detection” satisfactorily in partial fulfillment for the requirement of B.Sc.in
CSE. Bangladesh University of Business and Technology in the year 2021.

---------------------- ---------------------- ----------------------

Md. Mahedy Hasan Mahmudul Hasan Mahady Hasan
ID: 18191103048 ID: 18191103057 ID: 18191103061
Intake: 40 Intake: 40 Intake: 40

---------------------- ----------------------
Mohammad Tanvir Hossain Reza Momin Ali
ID: 17182103020 ID: 171812103011
Intake: 38 Intake: 38

--------------------------------

Project Supervisor

(MD. Shahiduzzaman)

Assistant Professor

Department of Computer Science and Engineering (CSE)

Bangladesh University of Business and Technology (BUBT)
 DEDICATION

Dedicated to our parents for all their love and inspiration.

 ACKNOWLEDGEMENTS

First of all, we are thankful and expressing our gratitude to Almighty Allah who offers us His
divine blessing, patience, mental and physical strength to complete this project work.

We are deeply indebted to our project supervisor MD. Shahiduzzaman, Assistant Professor,
Department of Computer Science and Engineering (CSE), Bangladesh University of Business and
Technology (BUBT). His scholarly guidance, important suggestions, work for going through our
drafts and correcting them, and generating courage from the beginning to the end of the research
work has made the completion of this project possible.

A very special gratitude goes out to all our friends for their support and help to implement our
works. The discussions with them on various topics of our works have been very helpful for us to
enrich our knowledge and conception regarding the work.

Last but not the least; we are highly grateful to our parents and family members for supporting us
spiritually throughout writing this report and our life in general.
 APPROVAL

This project report “Age and Gender Detection” Submitted by Md. Mahedy Hasan, Mahmudul
Hasan, Mahady Hasan, Mohammad Tanvir Hossain Reza and Momin Ali ID NO:
18191103048, 18191103057, 18191103061, 17182103020, 17182103011 Department of
Computer Science and Engineering (CSE), Bangladesh University of Business and Technology
(BUBT) under the supervision of MD. Shahiduzzaman, Assistant Professor, Department of
Computer Science and Engineering has been accepted as satisfactory for the partial fulfillment of
the requirement for the degree of Bachelor of Science (B.Sc. Eng.) in Computer Science and
Engineering and approved as to its style and contents.

--------------------------------

Supervisor:

MD. Shahiduzzaman

Assistant Professor
Department of Computer Science and Engineering (CSE)
Bangladesh University of Business and Technology (BUBT)
Mirpur-2, Dhaka-1216, Bangladesh

--------------------------------

Chairman:

Dr.Muhammad Firoz Mridha

Associate Professor & Chairman

Department of Computer Science and Engineering (CSE)
Bangladesh University of Business and Technology (BUBT)
Mirpur-2, Dhaka-1216, Bangladesh
© Copyright by Md. Mahedy Hasan (18191103048), Mahmudul Hasan (18191103057), Mahady
Hasan (18191103061), Mohammad Tanvir Hossain Reza (17182103020) and Momin Ali
(17182103011)

All Right Reserved

 Abbreviations
Synonyms and Acronyms Descriptions

CNN Convolutional Neural Network

ANN Artificial Neural Network

 Table of Content
Page No
Abstract iii

Declaration v
Certification vi
Acknowledgement vii
Dedication viii
Approval ix
Copyright x
Abbreviations xi

Chapter 01: Introduction 16

1.1 Introduction 16

1.2 Problem Statement 18

1.3 Problem Background 19
1.4 Research Objective 19
1.5 Motivation 19
1.6 Flow of Research 19
1.7 Significance of the Research 20
1.8 Research Contribution 21
1.9 Thesis Organization 21
1.10 Summery 21

Chapter 02: Literature Review 22-25

2.1 Introduction 22
2.2 Literature Review 22
2.3 Problem Analysis 24
2.4 Summary 25

Chapter 3: Proposed Model 26-31

3.1 Introduction 26

3.2 Feasibility Analysis 26

3.3 Requirement Analysis 27
3.4 Research Methodology 27
3.4.1 Architecture of our Model 28
3.4.2 DFD in our Model 30
3.4.3 UML Design in our Model 30
3.5 Design, Implementation and Simulation 30
3.6 Summary 31

Chapter 4: Implementation and Testing 32-38

4.1 Introduction 32
4.2 System Setup 33
4.3 Evaluation 34
4.4 Result and Discussion 35
4.5 Summary 38

Chapter 5: Conclusion 39-42

5.1 Conclusion 39
5.2 Future Work 39

Reference 40

Appendices 41

List of Tables xiv

List of Figures xiii

 List of Figures
Page No
Figure-1.1: Research Flow 19
Figure-3.1: Architecture of CNN Model 27
Figure-3.2: DFD in our Model 28
Figure-3.3: UML Design in our Model 29
Figure-4.1: System Architecture 32
Figure-4.2: Result Example 35
Figure-4.3: Result Example 36
 List of Tables

N/A
 Chapter 01
Introduction

3.4 Introduction

Age and gender, two of the key facial attributes, play a very foundational role in social interactions,
making age and gender estimation from a single face image an important task in intelligent
applications, such as access control, human-computer interaction, law enforcement, marketing
intelligence and visual surveillance, etc. The enhancement of raw images that are received from
the camera sources, from satellites, aircrafts and the pictures captured in day-to-day lives is called
image processing. The images have been processed through many different techniques and
calculations have been made

on the basis and analysis of the studies. There is a need to analyze and study the digitally formed
images. There are two main and very common steps followed for image processing. The
improvement of an image such that the resulting image is of greater quality and can be used by
other programs, is called image enhancement. The other technique is the most sought-after
technique used for extraction of information from an image. There is a division of the image into
a certain number of parts or objects so that the problem is solved. This process is called
segmentation. A neural network consists of many simple and similar compressing elements. It is
a system with inputs and outputs. There are a number of internal parameters called weights. An
artificial neural network is made of a set of processing elements which are also known as neurons
or nodes. These nodes are interconnected. Training in ANN is done through the track of the
examples. There are various such methods that fail to produce appropriate results. For each class,
an essential rule called the characteristic rule is generated. This set of rules is also called
differentiating rules. A systematic method which is used to train multilayer artificial neural
networks is known as back propagation. It is also considered as a gradient method where the
gradient of the error is evaluated by considering the weights of the given inputs. The detection of
the data available in the images is very important. The data that the image contains is to be changed
and modified for the detection purposes. There are various types of techniques involved for
detection as well as the removal of the problem. In a Facial detection technique: The expressions
that the faces contain hold a lot of information. Whenever a person interacts with the other person,
there is an involvement of a lot of expressions. The changing of expressions helps in calculating
certain parameters. Age estimation is a multi-class problem in which the years are classified into
classes. People with different ages have different facials, so it is difficult to gather the images.
Various age detection methods are used. The preprocessing is applied to the image. Features are
extracted from the neural network through the convolution network. Based on the trained models
the image is then classified to one of the age classes. Features are extracted from the images for
further processing. The features are processed further and sent to the training systems. The
databases provide a study of the features and help in completing the face detection for proving the
age detection of the person in the image. Age and gender assume essential parts in social activities.
Dialects hold distinctive greetings and grammar rules for men or women, and frequently diverse
vocabularies are utilized while tending to senior citizens compared to youngsters. In spite of the
essential parts these characteristics play in our everyday lives, the capacity to consequently assess
them precisely and dependably from face image is still a long way from addressing the
requirements of business applications. This is especially puzzling while considering late claims to
super-human capacities in the related errand of face recognition. .Past ways to deal with assessing
or ordering these properties from face images have depended on contrasts in facial feature
dimensions or “customized” face descriptors. Most have utilized characterization plans composed
especially for age or gender orientation estimation undertakings, including and others. Few of
these past strategies were intended to handle the numerous difficulties of unconstrained imaging
conditions . In addition, the machine learning strategies utilized by these frameworks did not
completely abuse the huge quantities of image cases and information accessible through the
Internet keeping in mind the end goal to enhance characterization capacities.In this paper we
endeavour to close the gap between automatic face recognition abilities and those of age and
gender classification techniques. To this end, we take after the fruitful sample set around late face
recognition frameworks: Face recognition systems portrayed in the most recent couple of years
have demonstrated that gigantic advancement can be made by the utilization of profound
convolutional neural networks (CNN) . We show comparative additions with basic system
engineering, composed by considering the somewhat constrained accessibility of precise age and
gender classification names in existing face information sets.

1.2 Problem Statement

Age and gender characteristics can be applied in retail for contextual advertising for particular
groups of customers. The reliability of the existing solutions remains insufficient for practical
application. Developing an age and gender recognition using convolutional neural networks with
classifier fusion methods

● To conduct a review of the existing solutions for image recognition by age and gender
with an emphasis on convolutional neural networks (CNNs)

● To describe the algorithm scheme of the proposed system

● To develop the age/gender recognition architecture, taking into account the classifier
fusion methods

● To analyze the accuracy of decision making on the basis of each aggregation method by
conducting experiments

1.3 Problem Background

Different methods have been proposed for extracting features from face images and to train the
system for identifying gender. We have experimented on different feature extraction methods
and been adopted for our proposed system. In the following subsections we will briefly describe
some necessary topics related to our proposed system.

1.4 Research Objectives

The objective of this study is as follows research:
● To Identify Age of Human
● To Identify Gender of Human
● To Achieve Better Accuracy

1.5 Motivation
The motivation behind this project was to build an application for age and gender classification
using a model that provides better accuracy. Many models are focusing on real life detection in-
the-wild estimation. That doesn’t provide better accuracy. In this project we will focus on CNN
methods. That will help us to achieve good accuracy.

1.6 Flow of Research

The research work is being developed in several steps. Firstly we analyzed the research topic then
studied the basic theory of CNN and OpenCV. Then we investigated the application of CNN and
OpenCV. We investigated the lack of present architectures and motivated them to build a new
architecture based on CNN architecture. Figure 1.1 illustrates the overall steps to the research
procedure in the following diagram.
 Start

Understanding and experimenting with different areas of CNN

and OpenCV, Age and Gender prediction methodology

Literature review on Age and Gender Detection

Choose an existing dataset from kaggle

Classify Image using CNN and OpenCV

Make decision about age and gender

Provide result and find the conclusion

End

Figure 1.1 Research Flow

1.7 Significance of the research

We observe that most of the Age and Gender Detection datasets on the web are built on random
pictures of humans. However, to analyse Age and Gender, we needed a dataset with a huge
amount of pictures. Therefore we use CNN approaches to identify Age and Gender. Nowadays,
automated markets/shopping malls require Age and Gender data. Age and Gender Detection can
help to detect age and gender of people. It can help to detect age and gender in any field of our
daily life.
1.8 Research Contribution

The overall contribution of the research work are :

● The project requires study of research on CNN and database learning.

● Data collection and information.
● Engineering design and development.
● Knowledge of software engineering and image processing.
● In this project where high definition images will be used, but the system will be able to
predict even low quality images.

1.9 Thesis Organization

The thesis work is rganized as follows.

Chapter 2 highlights the background and literature review on the field of Age and Gender
Detection. Chapter 3 contains Age and Gender Detection proposed architecture and a detailed
walk-through of the overall procedures. Chapter 4 includes the details of the tests and evaluations
performed to evaluate our proposed architecture. Chapter 5 explains the Standards, Impacts,
Ethics, Challenges, the Constraints, Timeline, and Gantt Chart. Finally, Chapter 6 contains the
overall conclusion of our thesis work.

1.10 Summary

This chapter includes a broad overview of the problem that we aimed explicitly at our research
work’s objectives, the background, and the research work’s motivation. This chapter also
illustrates the overall steps on which we carried out our research work.
 Chapter 02

Literature Review

2.1 Introduction
This section succinctly reviews the associated works of age and gender estimation. Although a
significant amount of literature is already available related to these topics, we try to provide a
superficial outline in what way these tasks were approached earlier by other researchers.

2.2 Literature Review

Age estimation is a long-studied research topic among computer vision researchers. Most of the
researchers considered human age estimation as either a classification or regression problem. In
the case of age classification, age is coupled with a specific range or age group. On the other hand,
age regression is a single value estimated for a person. However, it is very challenging to estimate
an exact age due to diversity in the aging process across different ages. Furthermore, for accurate
age estimation, the model needs a huge amount of correctly labeled face data. Many of the early
age estimation methods used hand-crafted facial features in the constrained imaging conditions. A
survey of such methods was reported in and a recent survey of age estimation including all
approaches of the last decades can be found in . A method that extracts the geometric features from
the face and calculates ratios among the facial features to estimate the age is presented in [28].
Initially, the face wrinkles are detected and localized, then the size and distances of the facial
features are measured and finally the face is classified into different age categories. A similar
approach as presented in [28] is proposed in [29] by modeling growth-related shape variations
observed in human faces considering anthropometric evidence. This work was limited to a certain
age. The above-mentioned methods are unsuitable for images in-the-wild due to the necessity of
accurate localization of facial features. A couple of subspace methods were introduced, where
aging features were extracted from an aging pattern representative subspace and a robust
regression was used to predict the face ages. Although the aforementioned methods achieved
excellent performances compared to previous cutting-edge methods, some limitations are exposed
by these algorithms. Their system worked well only with frontal and properly aligned images. The
algorithms proposed by these researchers are not well suited for practical applications where the
input images might be collected in an unconstrained environment.
There are some methods where face images are represented using spatially localized facial patches.
In , the patch distribution was represented by exploiting probabilistic Gaussian mixture models .
A robust descriptor was used in place of pixel patches in . Later on, the Hidden-Markov-Model
was introduced instead of the Gaussian mixture model (GMM) for representation of face patch
distribution . A significant amount of research employed different robust image descriptors as an
alternative to the local image intensity patch for the age estimation task. Gao et al. used Gabor
filters along with a Fuzzy-LDA classifier where one face belongs to multiple age classes.
Similarly, Biologically-Inspired-Features in , and Local Binary Patterns (LBP) were presented in
. The age estimation problem was considered as a regression problem by Fu et al. or as a
classification problem using a quadratic function, shortest distance, and neural network-based
classifiers in . The popular regression techniques reported by the researchers are Support Vector
Regression (SVR) , Partial Least Squares (PLS) , Canonical Correlation Analysis (CCA) .
Accordingly, Nearest Neighbor (NN) and Support Vector Machines (SVM) are the most used
classification techniques. Next, we choose a couple of real age estimation methods to describe
those that are most related to our suggested method. Guo et al. presented a learning scheme to
draw aging features named manifold learning and utilized SVRs with local adjustment for age
prediction, Han et al. extracted features using boosting algorithms and formed a hierarchical
approach for classification between age group and regression inside a group (DIF). Geng et al.
introduced an aging pattern subspace (AGES) wherefrom features were extracted and performed
regression for age estimation. Zhang and Yeung handled age estimation as a multi-task problem
based on the warped Gaussian process (MTWGP) where common features were shared among the
tasks. Chen et al. Introduced a mapping between the cumulative attribute space and low-level
sparse features for age regression. Chang et al. Formed the age labels into binary groups that
formed subproblems and imposed a cost on each subproblem. Thus, they ranked the ordinal
hyperplanes based on classification cost for age estimation, while Guo and Mu used a canonical
correlation analysis and partial least squares founded methods to perform feature projection and
estimate human traits jointly. Recently, the biologically inspired CNN models were successfully
deployed for the age estimation task. Yi et al. [51] deployed a multiscale CNN. Wang et al. [52]
used features from the intermediate layer of CNN rather than top layer features and performed
manifold learning. Rothe et al. Incorporated a deep CNN for extracting features and real age
regression as an estimate using SVR. In, they used a deeply learned CNN model from large in-the-
wild image data and performed age regression through classification. A hybrid system was
introduced, where the CNNs were used for face feature extraction and an extreme learning machine
(ELM) for the classification task. A lightweight CNN network with mixed attention mechanism
for low end devices was proposed in , where the output layer was fused by classification and
regression approach. Another multi-task learning approach merging classification and regression
concepts to fit the age regression model with heterogeneous data with the help of two different
techniques for partitioning data towards classification was proposed in . To resolve the problem of
data disparity and ensure the generality of the model, a very recent method is proposed by Kim et
al. Where a cycle generative adversarial network-based race and age image transformation method
is used to generate sufficient data for each distribution. All of the aforementioned CNN-based
systems are evaluated on the basis of the common dataset Morph for age estimation. To the best
of our knowledge, demonstrate state-of-the-art results. A lot of progress has been achieved in the
gender estimation topic, but it is still a challenging problem in the real-world environment. The
literature about gender estimation comes under the umbrella of the authors of . Here, we will
discuss some of those methods where the well-known classifiers are used for gender estimation.
As one of the very early methods, the authors used a fully connected two-layer neural network that
learned from a limited number of near-frontal face images for gender classification. SVM
classifiers were directly applied to image intensities. Similarly, AdaBoost was introduced instead
of the SVM classifier by keeping the same working pipeline . Later on, a viewpoint-invariant
model for age and gender estimation was suggested by Toews and Arbel which is robust to local
scale rotations. A combination of human knowledge and a gait information-based gender
classification system was provided by Yu et al. An unconstrained face image benchmark for gender
classification along with a high classification accuracy was presented in. Khan et al. Formed a
semantic pyramid by extracting features from the full and upper body together with face regions
from the image to recognize the gender and action. This method does not depend on the annotation
value of a person’s face and upper body to extract semantic features. In , the authors proposed a
model where the first name of a person is used as a special feature and associates a name with
facial appearance to recognize the gender of that person. At the same time, the authors showed that
their method achieved higher accuracy in the task of gender recognition and demonstrated the
potential in the use of face verification. In recent times, a generic framework for age and gender
estimation was proposed in, where a hierarchical estimator was modeled based on extracted
biologically inspired features. Besides, this method was formed to detect low-quality images due
to a poor image background. The above efforts regarding gender estimation contributed a lot in
this research area. However, the lion’s share of these methods is only suitable for the applications
with constraint images or have higher computational costs. Recently, a deep CNN-based approach
was presented in. It was pretrained with a huge unconstrained dataset and then fine-tuned on two
other datasets to achieve a very good accuracy in the gender estimation task. Although, their
method states a high accuracy, a lot of pretraining is required prior to evaluating the system. In our
paper, we propose a system that will work comparatively well with unconstrained imaging
conditions.

2.3 Problem Analysis

We find CNN, SVM, KNN algorithms from the research paper where they did Age and Gender
Classification. The researcher gets better performance using Convolutional Neural Network
(CNN) architecture that is based on ANN. ANN is very popular in classification. ANN techniques
are more suitable than other techniques.

2.4 Summary
This chapter includes an overview of the problem that we specifically target, the objectives of
our thesis work along the motivation of the output of the thesis work. This section also
elaborates on the overall steps on which we carried out our thesis work.
 Chapter 03
Proposed Model

3.1 Introduction

This section presents the proposed deeply learned classifiers for age group and gender
classification of unfiltered real-life face images, an extension of our conference paper in .

The approach as presented in Algorithm 1 requires image preprocessing (face detection, landmark
detection, and face alignment) stage that preprocess and prepare the face images before they are
input into the proposed network. Therefore, our solution is divided into three major steps: image
preprocessing, features learning, and classification itself.

Intelligent age and gender classifiers tackle the classification task under unfiltered real-world
settings. Most of those face images are not aligned and non frontal and also with different degrees
of variations in pose, appearance, lighting, and background conditions. Therefore, those in-the-
wild face images need first to be detected, then aligned, and, finally, used as input for the
classifiers.

3.2 Feasibility Analysis

1. Technical Feasibility –we use this technology in this program
● Pycharm
● Python IDE
● PIP Installer
 2. Economic Feasibility –
There is no economic feasibility required because we find every software we use in
this program as free. And also we have our required hardware available.

3.3 Requirement Analysis

The proposed model includes:

Very fast computing device.

Pycharm/Kaggle

Open-source library function to calculate performance.

3.4 Research Methodology

In this section, we discuss our proposed CNN model overall process of how it works. This
section divided 4 steps other sub-steps that show the input and output process of our proposed
model. Each step shows detailed works and proper explanation.
3.4.1 Architecture of our Model

Figure 3.1: Architecture of our CNN Model

3.4.2 DFD in our model

Figure 3.2: DFD in our Model

3.4.3 UML Design in our Model

Figure 3.3: UML Design

3.5 Design, Implementation, and Simulation

Figure 3.1 explains the details. All steps of this model have implemented Python
language and the details work throw platform is google collaborator. The Convolutional neural
network models are implemented using python. Also, for additional calculation, and support,
Scikit, and Numpy-pandas are used. The visual evaluation reports were generated using

Matplotlib in python.
3.6 Summary

The proposed approach for estimating the real age and gender of a person demonstrates state-of-
the-art results for the Morph dataset in anticipation of real age and the Adience dataset for gender
estimation. We surpass the age estimation results marginally compared to the state-of-the-art DEX
[24] method by lessening the data sparsity problem that exists in their approach. Our approach for
the age estimation reduces the mean average error reported in the SOA method, i.e., the DEX fine-
tuned for the IMDB-WIKI dataset and without fine-tuning by 0.26 years and 0.57 years,
respectively. The system reveals that pre-training on a balanced dataset along with sufficient
training data boosts the system performance reasonably. Our target application is a real-time
system. Therefore, we incorporate the batch normalization layer before the non-linear RELU layer
in the model to achieve a quicker convergence. In Table 3, it is empirically demonstrated that
training the CNN model using a balanced dataset with batch normalization takes the lead in the
result table. For the gender estimation, we used the ImageNet pre-trained model as the age
estimation task without pre-training on the IMDB-WIKI dataset. We fine-tune and evaluate the
model with the heterogeneous Audience data set and achieve SOA results. We achieve 5% more
accuracy than the state-of-the-art approach [81]. In future research, we will try to enrich the
balanced dataset because the amount of data in every class is not sufficient enough to train a very
deep model and secure an enviable performance for a complex problem like real age estimation.
We will try to devise an optimal network for the age and gender estimation from masked face
images since smart store customers must wear a face mask amid the pandemic situation.
 Chapter 04
Implementation and Testing

4.1 Introduction

This technique involves computation of a set of geometrical features such as nose width and
length, mouth position and chin shape, etc. from the picture of the face we want to recognize. This
set of features is then matched with the features of known individuals. A suitable metric such as
Euclidean distance (finding the closest vector) can be used to find the closest match. Most
pioneering work in face recognition was done using geometric features (Kanade, 1973), although
Craw et al. (1987) did relatively recent work in this area. The advantage of using geometrical
features as a basis for face recognition is that recognition is possible even at very low resolutions
and with noisy images (images with many disorderly pixel intensities). Although the face cannot
be viewed in detail its overall geometrical configuration can be extracted for face recognition. The
technique's main disadvantage is that automated extraction of the facial geometrical features is
very hard. Automated geometrical feature extraction based recognition is also very sensitive to the
scaling and rotation of a face in the image plane (Brunelli and Poggio, 1993). This is apparent
when we examine Kanade's(1973) results where he reported a recognition rate of between 45-75
% with a database of only 20 people. However if these features are extracted manually as in
Goldstein et al. (1971), and Kaya and Kobayashi (1972) satisfactory results may be obtained.
4.2 System Setup
Automated age and gender estimation of the human face involves detecting, tracking and
normalizing the face in an image sequence. The features extracted from the face image are used
to estimate age and gender, and a face tracking method is employed to measure stay time and to
count people. Figure 1 illustrates the system architecture used in this study.

Figure 4.1: System architecture for age and gender estimation

4.3 Evaluation

We propose a model that uses CNN architecture to predict the age group and gender of human’s
faces from unfiltered real-world environments. The novel CNN approach addresses the age and
gender labels as a set of discrete annotations and trains the classifiers that predict the human’s age
group and gender.

We design a quality and robust image preprocessing algorithm that prepares and preprocess the
unfiltered images for the CNN model and this greatly has a very strong impact on the performance
accuracy of our age and gender classifiers.

We demonstrate that pretraining on large-scale datasets allows an effective training of our age and
gender CNN model which enable the classifiers to generalize on the test images and then avoid
overfitting

Finally, OIU-Adience benchmark is used to evaluate the performance of our novel CNN model,
and despite the very challenging nature of the images in the dataset, our approach produces
significant improvements in age group and gender classification accuracy over the state-of-the-art
methods; the result can satisfy the requirements of several real-world applications.
4.4 Results and Discussion

At the beginning of this section, we discuss the implementation details of the age and gender
estimation task. Next, we introduce the datasets used for both tasks. Subsequently, we report both
quantitative and qualitative results observed following the experiments. Finally, we discuss the
results at the end of this section. We trained our deployed CNN model separately based on the age
and gender classification task. We handle both tasks as a classification approach. The models are
trained using the deep learning framework.In the case of age classification, we calculate the
expected value from the Softmax probabilities belonging to output neurons, similar to what was
considered in. We train the model with a balanced IMDB-WIKI dataset and introduce the batch
normalization layer to reduce the training time. We consider every age as an individual class that
ranges from 0 to 100. For all experiments regarding age estimation, the CNN is initialized with the
weights trained on ImageNet . This pre-trained model is then further trained on the IMDB-WIKI
image dataset for classification with 101 output neurons. Finally, the CNN is fine-tuned on the test
dataset. For the gender classification, we report the gender class of the neuron carries the highest
probability. We first deployed the pre-trained model trained on ImageNet. In the next step, we
fine-tuned the pre-trained model with the real-world face image dataset added with two output
neurons and the performance is reported from the test split of the Adience dataset. The training set
consists of 80% of the images from the dataset and 20% is reserved for the testing. Further 90%
of images from the training set are used for learning the weights and the rest of the images are used
as validation sets during the training phase. Every experiment begins in conjunction with pre-
trained ImageNet weights. When fine-tuning the pre-trained network with a smaller dataset, the
learning rate of 0.001 remains fixed except for the last layer. The last layer weights are initialized
randomly as the number of output neurons are changing. We used the Adam optimizer with the
setting of momentum 0.9 and a weight decay rate of 5 × 10−4 . We adjust the learning rate by a
factor of 10 after every 30 epochs.
In this paper, we use four different datasets for real age and gender estimation. We first introduce
the datasets with a description of their specifications. Figure 7 represents exemplar images for each
dataset used for the age estimation experiment and sample images for gender prediction are
presented in Figure 8. Table 4 shows the size of each dataset with its properties. For the age
classification, we trained our model with the IMDBWIKI dataset and fine-tuned it with the
MORPH dataset. We evaluate our age estimation model with the MORPH dataset. In the gender
estimation task, we train and evaluate our model with the dataset.Figure-Exemplar images with
real (biological) age from each dataset used in the age estimation experiment.

Figure 4.2: Result Example

Figure 4.3: Result Example
4.5 Summary
The proposed approach for estimating the real age and gender of a person demonstrates state-of-
the-art results for the Morph dataset in anticipation of real age and the dataset for gender
estimation. We surpass the age estimation results marginally compared to the state-of-the-art DEX
method by lessening the data sparsity problem that exists in their approach. Our approach for the
age estimation reduces the mean average error reported in the SOA method, i.e., the DEX fine-
tuned for the IMDB-WIKI dataset and without fine-tuning by 0.26 years and 0.57 years,
respectively. The system reveals that pre-training on a balanced dataset along with sufficient
training data boosts the system performance reasonably. Our target application is a real-time
system. Therefore, we incorporate the batch normalization layer before the non-linear RELU layer
in the model to achieve a quicker convergence. In Table 3, it is empirically demonstrated that
training the CNN model using a balanced dataset with batch normalization takes the lead in the
result table. For the gender estimation, we used the ImageNet pre-trained model as the age
estimation task without pre-training on the IMDB-WIKI dataset. We fine-tune and evaluate the
model with the heterogeneous dataset and achieve SOA results. We achieve 5% more accuracy
than the state-of-the-art approach . In future research, we will try to enrich the balanced dataset
because the amount of data in every class is not sufficient enough to train a very deep model and
secure an enviable performance for a complex problem like real age estimation. We will try to
devise an optimal network for the age and gender estimation from masked face images since smart
store customers must wear a face mask amid the pandemic situation.
 Chapter 05
Conclusion

5.1 Conclusions

In this paper, we propose an automatic age (biological) and gender estimation system for the
promising smart store enterprise which is a modern venture in the retail industry. This automated
system can extract the human demographics necessary to provide the customer a very good
shopping experience that results in a boost of offline smart store sales. In addition, this enterprise
solution can ease the shopping process and shorten the shopping time for the consumers. Although
recent methods show their potential for the problem of age and gender estimation, the best works
focused on constrained image benchmarks. As a result, these methods are not robust enough for
the application involving real-world images. Most recently, some of the researchers learned to
apply their model utilizing unconstrained image datasets, but these models are biased for the early
and middle adulthood classes due to image sparsity problems in the dataset. In our paper, we
resolve the data sparsity problem that exists in the state-of-the-art real age estimation approach
DEX by constructing class-wise data parity and incorporate batch normalization concepts that
jointly improve the previous results marginally. We follow the same setup for the gender
estimation task and achieve substantially improved results over SOA methods regarding this task.

Author Contributions: Conceptualization, M.M.I. and J.-H.B.; methodology, M.M.I.; software,

M.M.I.; validation, M.M.I. and J.-H.B.; formal analysis, M.M.I.; investigation, M.M.I; data
curation, M.M.I.; writing—original draft preparation, M.M.I.; writing—review and editing, J.-
H.B. and M.M.I.; visualization, M.M.I.; supervision, J.-H.B. All authors have read and agreed to
the published version of the manuscript.

Funding: This research is supported by the GRRC program of Gyeonggi province [GRRC
Aviation 2017-B04, Development of Intelligent Interactive Media and Space Convergence
Application System].

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Data Availability Statement: The authors have used publicly archived IMDB-WIKI and
Adience dataset for the experiments. The IMDB-WIKI dataset is available in . The Audience
dataset is available in [9].
Acknowledgments: We would like to acknowledge Korea Aerospace University with much
appreciation for its ongoing support to our research. We are thankful to authors who make their
datasets publicly available to pave the way to research in this area.

Conflicts of Interest: The authors declare no conflict of interest.

5.2 Future work

In our system we have used the modern web technologies to make our system fast, convenient and
efficient for all of the personnel mentioned.Due to time and cost constraint it was not possible to
fulfill all requirements and functionalities those were planned. But in future these planned
functionalities and more improvement will be possible to pursue. The functionalities to be
implemented areThus, there is a place for some future work such as:

● Different classifications models can be used at a time to improve the performance of the
segmentation.

●To reduce the complexity of the algorithm,it’s better to reduce the number of hypothesis to
function the algorithm faster.

●To reduce the computation time, better filters are to be used to eliminate the unnecessary
segmentation hypothesis.

●In future we want to implement distributed database.

●We will develop an android and iOS platform application.

●We will try to make easier and more reliable and handier user interface so that users can easily
operate this application.

We also want to modify our application for the Future Work, we hope this work will help us in our
future work. Our system is developed based on demand of user’s satisfaction and facilities. In our
system we have used the modern web technologies to make our system fast, convenient and
efficient for all of the personnel mentioned. Due to time and cost constraint it was not possible to
fulfill all requirements and functionalities those were planned.
Reference

1. Ricanek, K.; Tesafaye, T. Morph: A Longitudinal Image Database of Normal Adult Age-
Progression. In Proceedings of the Seventh International Conference on Automatic Face and
Gesture Recognition, Southampton, UK, 10–12 April 2006; pp. 341–345.

2. Geng, X.; Zhou, Z.-H.; Zhang, Y.; Li, G.; Dai, H. Learning from Facial Aging Patterns for
Automatic Age Estimation. In Proceedings of the ACM International Conference on Multimedia,
Santa Barbara, CA, USA, 23–27 October 2006; pp. 307–316.

3. Minear, M.; Park, D.C. A lifespan database of adult facial stimuli. Behav. Res. Methods
Instrum. Comput. 2004, 36, 630–633.

4. Choi, S.E.; Lee, Y.J.; Lee, S.J.; Park, K.R.; Kim, J. Age Estimation Using a Hierarchical
Classifier Based on Global and Local Facial Features. Pattern Recognit. 2011, 44, 1262–1281.

5. Song, Z.; Ni, B.; Guo, D.; Sim, T.; Yan, S. Learning Universal Multi-View Age Estimator
Using Video Context. In Proceedings of the International Conference on Computer Vision,
Barcelona, Spain, 6–13 November 2011; pp. 241–248.

6. A. Kumar and F. Shaik,” Importance of Image Processing”, 2016.

7. Chenjing Yan, Congyan Lang, Tao Wang, Xuetao Du, and Chen Zhang,” Age Estimation
Based on Convolutional Neural Network”, 2014 Springer International Publishing Switzerland.

8. Hang Qi and Liqing Zhang,” Age Classification System with ICA Based Local Facial
Features”, 2009 Springer-Verlag Berlin Heidelberg.

9. Eran Eidinger, Roee Enbar, and Tal Hassner,” Age and Gender Estimation of Unfiltered
Faces”, 2014 IEEE.

10. Aditya K. Saxena, Shweta Sharma and Vijay K. Chaurasiya,” Neural Network based Human
Age-group Estimation in Curvelet Domain”, 2015 The Authors Published by Elsevier.

11. J. Chen, S. Shan, C. He, G. Zhao, M. Pietikainen, X. Chen, and W. Gao. Wld: “A robust
local image descriptor”, Trans.Pattern Anal. Mach. Intell.,

12. S. E. Choi, Y. J. Lee, S. J. Lee, K. R. Park, and J. Kim. “Age estimation using a hierarchical
classifier based on global and local facial features”, Pattern Recognition, 44(6):1262–1281, 2011

13. T. F. Cootes, G. J. Edwards, and C. J. Taylor. “Active appearance models”, In European

Conf. Comput. Vision, pages 484–498. Springer, 1998.

14. C. Cortes and V. Vapnik. “Support-vector networks”, Machinelearning, 20(3):273–297,

1995.
15. E. Eidinger, R. Enbar, and T. Hassner. “Age and gender estimation of unfiltered faces”,
Trans. on Inform.Forensics andSecurity, 9(12), 2014.

16. Saragih, J.M.; Lucey, S.; Cohn, J.F. Deformable model fitting by regularized landmark mean-
shift. Int. J. Comput. Vis. 2011, 91, 200–215.
Appl. Sci. 2021, 11, 4549 21 of 23

17. Xiong, X.; De la Torre, F. Supervised Descent Method and Its Applications to Face
Alignment. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition
(CVPR), Portland, OR, USA, 23–28 June 2013; pp. 532–539.

18. Davies, G.; Ellis, H.; Shepherd, J. Perceiving and remembering faces. Am. J. Psychol. 1983,
96, 151–154.

19. Sinha, P.; Balas, B.; Ostrovsky, Y.; Russell, R. Face recognition by humans: Nineteen results
all computer vision researchers should know about. Proc. IEEE 2006, 94, 1948–1962.

20. Gross, R.; Matthews, I.; Baker, S. Generic vs. person specific active appearance models.
Image Vis. Comput. 2005, 23, 1080–1093.
Appendices

import cv2
import math
import argparse

def highlightFace(net, frame, conf_threshold=0.7):

frameOpencvDnn=frame.copy()
frameHeight=frameOpencvDnn.shape[0]
frameWidth=frameOpencvDnn.shape[1]
blob=cv2.dnn.blobFromImage(frameOpencvDnn, 1.0, (300, 300), [104, 117, 123], True,
False)

net.setInput(blob)
detections=net.forward()
faceBoxes=[]
for i in range(detections.shape[2]):
confidence=detections[0,0,i,2]
if confidence>conf_threshold:
x1=int(detections[0,0,i,3]*frameWidth)
y1=int(detections[0,0,i,4]*frameHeight)
x2=int(detections[0,0,i,5]*frameWidth)
y2=int(detections[0,0,i,6]*frameHeight)
faceBoxes.append([x1,y1,x2,y2])
cv2.rectangle(frameOpencvDnn, (x1,y1), (x2,y2), (0,255,0), int(round(frameHeight/150)), 8)
return frameOpencvDnn,faceBoxes

parser=argparse.ArgumentParser()

parser.add_argument('--image')

args=parser.parse_args()

faceProto="opencv_face_detector.pbtxt"
faceModel="opencv_face_detector_uint8.pb"
ageProto="age_deploy.prototxt"
ageModel="age_net.caffemodel"
genderProto="gender_deploy.prototxt"
genderModel="gender_net.caffemodel"

MODEL_MEAN_VALUES=(78.4263377603, 87.7689143744, 114.895847746)

ageList=['(0-2)', '(4-6)', '(8-12)', '(15-20)', '(25-32)', '(38-43)', '(48-53)', '(60-100)']
genderList=['Male','Female']
faceNet=cv2.dnn.readNet(faceModel,faceProto)
ageNet=cv2.dnn.readNet(ageModel,ageProto)
genderNet=cv2.dnn.readNet(genderModel,genderProto)

video=cv2.VideoCapture(args.image if args.image else 0)

padding=20
while cv2.waitKey(1)<0:
hasFrame,frame=video.read()
if not hasFrame:
cv2.waitKey()
break

resultImg,faceBoxes=highlightFace(faceNet,frame)
if not faceBoxes:
print("No face detected")

for faceBox in faceBoxes:

face=frame[max(0,faceBox[1]-padding):
min(faceBox[3]+padding,frame.shape[0]-1),max(0,faceBox[0]-padding)
:min(faceBox[2]+padding, frame.shape[1]-1)]

blob=cv2.dnn.blobFromImage(face, 1.0, (227,227), MODEL_MEAN_VALUES,

swapRB=False)
genderNet.setInput(blob)
genderPreds=genderNet.forward()
gender=genderList[genderPreds[0].argmax()]
print(f'Gender: {gender}')

ageNet.setInput(blob)
agePreds=ageNet.forward()
age=ageList[agePreds[0].argmax()]
print(f'Age: {age[1:-1]} years')

cv2.putText(resultImg, f'{gender}, {age}', (faceBox[0], faceBox[1]-10),

cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0,255,255), 2, cv2.LINE_AA)
cv2.imshow("Detecting age and gender", resultImg)