A REPORT OF SEMESTER TRAINING (BTCS 801-18)

at

Website blocking
BACHELOR OF TECHNOLOGY

Department of Computer Science and Engineering

CHANDIGARH ENGINEERING COLLEGE JHANJERI, MOHALI

In partial fulfillment of the requirements for the award of the Degree of

Bachelor of Technology in Computer Science & Engineering

SUBMITTED BY: Under the Guidance of

Name: Ashu Kumar Er, Himani Ma’am
Roll no: 2026940

FEB 2024

Affiliated to I.K Gujral Punjab Technical University, Jalandhar

CERTIFICATE BY COMPANY
 Page I

DECLARATION

I “ASHU KUMAR” hereby declare that I have undertaken Semester Training at “U&M
Overseas Pvt. Ltd.” during a period from January 2024 to May 2024 in partial fulfillment
of requirements for the award of degree of B.Tech (Computer Science and Engineering) at
Chandigarh Engineering College Jhanjeri, Mohali. The work which is being presented in
the training report submitted to Department of Computer Science and Engineering at
Chandigarh Engineering College Jhanjeri, Mohali is an authentic record of training work.

ASHU KUMAR Er. Himani Ma’am

Univ. Roll No. 2026940 Ass. Professor(CSE) , CGC Jhanjeri
Semester: 8th

Signature of the Head of Department

(With Stamp)
 Page II

ACKNOWLEDGEMENT

The internship opportunity I had with SysCloud was a great chance for learning and
professional development. Therefore, I consider myself as a very lucky individual as I was
provided with an opportunity to be a part of it. I am also grateful for having a chance to
meet so many wonderful people and professionals who led me though this internship
period.
Bearing in mind previous I am using this opportunity to express my deepest gratitude and
special thanks to the CEO of SysCloud Mr. Vijay Krishna who despite being extraordinarily
busy with her/his duties, took time out to hear, guide and keep me on the correct path and
allowing me to carry out my training at their esteemed organization.
I express my deepest thanks to Pinklin Devamalar, HR for taking part in useful decision &
giving necessary advices and guidance and arranged all facilities to make life easier. I choose
this moment to acknowledge his/her contribution gratefully.
It is my radiant sentiment to place on record my best regards, deepest sense of gratitude to
Mr. Palani, Managing Director, for his careful and precious guidance which were extremely
valuable for my study both theoretically and practically.
I perceive this opportunity as a big milestone in my career development. I will strive to use
gained skills and knowledge in the best possible way, and I will continue to work on their
improvement, to attain desired career objectives. Hope to continue cooperation with all of
you in the future,
Sincerely,
SATISH CHANDRA

Page III
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

TABLE OF CONTENTS

PARTICULARS PAGE NO

Title Page I

Declaration by the Candidate II

Acknowledgement III

Table of Contents 1-2

Abstract 3

CHAPTER 1 INTRODUCTION 4-6

CHAPTER 2 REVIEW OF LITERATURE 7-14

CHAPTER 3 PROBLEM DEFINITION AND OBJECTIVES 15-16

3.1 Problem Formulation
3.2 Objective
CHAPTER 4 DESIGN AND IMPLEMENTATION 17-29
4.1 Design
4.2 Implementation
4.2.1 Model Training
4.2.2 Data preprocessing
4.2.3 Training process
4.2.4 Network
4.2.5 Evaluation of best training model

Page|1
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

TABLE OF CONTENTS

CHAPTER 5 SYSTEM REQUIREMENT ANALYSIS 30-51

5.1 Hardware Requirement

5.2 Network Requirement

CHAPTER 6 SOFTWARE REQUIREMENT ANALYSIS 52-53

CHAPTER 7 RESULT AND DISCUSSION 54

CHAPTER 8 CONCLUSION AND FUTURE SCOPE 55-57

8.1 Possible Future Scope

REFERENCE 58-59

Page|2
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

ABSTRACT

Compared with online advertising industry, there is an even faster increase of ad blocker usage, which

influence badly on publishers' and advertisers' business. Thus more and more companies initialize their

counter-ad blocking strategies, in which customers choose to either disable their ad blockers or leave without

seeing the content. There are also companies which abandon their counter-ad blocking strategies after

conducting them for a while due to insufficient understanding of users' ad blocking behavior. In this study,

we employed a quasi-experiment framework and collected a large-size data with the cooperation with Forbes

Media. We aim to identify factors influencing ad blocker usage. Furthermore, we will model the interaction

effects among user profile, online behavior patterns, device features on ad blocker usage propensity. Our

study contributes the literature of understanding ad blocker usage by evaluating those principles using big

amount of real-world data.

Page|3
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

INTRODUCTION

Now days the role of online marketing tools is increasing in e-business. Describing the ad blocking

software usage factors for future web advertisement strategy development is an important issue in

the context of online marketing tool effectiveness. The changes in online advertisements, native

advertisements and new versions of old “conventional advertisements,” depend on the behavior of

internet users, especially the marketing strategy of targeting users in e-business. This article is the result of

research study that was performed using the following methodology Ad Blocking and Counter-Ad Blocking

Twenty-third Americas Conference on Information Systems, Boston, 2017 1 Ad Blocking and Counter-Ad

Blocking: Analysis of Online Ad Blocker Usage Emergent Research Forum papers Shuai Zhao New Jersey

Institute of Technology [email protected] Chong Wang New Jersey Institute of Technology [email protected]

Achir Kalra Forbes Media [email protected] Leon Vaks New Jersey Institute of Technology

[email protected] Cristian Borcea New Jersey Institute of Technology [email protected] Yi Chen New

Jersey Institute of Technology [email protected] Abstract Compared with online advertising industry, there

is an even faster increase of ad blocker usage, which influence badly on publishers’ and advertisers’

business. Thus more and more companies initialize their counter-ad blocking strategies, in which

customers choose to either disable their ad blockers or leave without seeing the content. There are also

companies which abandon their counter-ad blocking strategies after conducting them for a while due to

insufficient understanding of users’ ad blocking behavior. In this study, we employed a quasi-experiment

framework and collected a large-size data with the cooperation with Forbes Media. We aim to identify

factors influencing ad blocker usage. Furthermore, we will model the interaction effects among user

profile, online behavior patterns,

Page|4
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

device features on ad blocker usage propensity. Our study contributes the literature of understanding

ad blocker usage by evaluating those principles using big amount of real-world data. Keywords online

advertising, big data analytics, ad avoidance, ad blocker Introduction Digital technologies and the Internet

have dramatically changed the content publishing industry whose major product is information goods.

Most content in the internet is free to read and the publishers make profits through digital advertising.

There is a striking growth of online advertising industry. According to reports of Interactive Advertising

Bureau (IAB), online display advertising has emerged as the largest medium after surpassing TV

broadcast advertising since 2013 (IAB 2014). The revenue of digital advertising in 2015 reached to

59.6 billion dollars in United States, 20.4% higher than in 2014 (IAB 2016). Online advertising aims at

increasing sales by making viewers to notice their advertised products when they are viewing the

publishers’ content, so as to entice them to buying the products (Danaher and Mullarkey 2003).

Typically, there are three basic stakeholders in online advertising market, i.e., readers, publishers and

advertisers. A publisher integrates ads into its online content. An advertiser provides the ads and pay

publishers for displaying the ads. Readers are the consumers of the content and the ads. However, ad

blockers have gained wide usage rapidly. An ad blocker is a tool, most likely a browser plugin, to remove

ads while a user is reading online content. As a result, the advertisers fail to make the marketing

via online ads and publisher suffers from the decrease of online ads revenue. According to a 2015 report

by PageFair on ad blocking, the number of ad blocker users has increased steadily by 41% year over

year and ad blocking is estimated to strip about 1/3 of the revenue of digital advertising in 2015.Digital

technologies and the Internet have dramatically changed the content publishing industry whose major

product is information goods. Most content in the internet is free to read and the publishers make profits

through

Page|5
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

digital advertising. There is a striking growth of online advertising industry. According to reports of

Interactive Advertising Bureau (IAB), online display advertising has emerged as the largest medium

after surpassing TV broadcast advertising since 2013 (IAB 2014). The revenue of digital advertising

advertising aims at increasing sales by making viewers to notice their advertised products when they

are viewing the publishers’ content, so as to entice them to buying the products (Danaher and

Mullarkey 2003). Typically, there are three basic stakeholders in online advertising market, i.e., readers,

publishers and advertisers. A publisher integrates ads into its online content. An advertiser provides the

ads and pay publishers for displaying the ads. Readers are the consumers of the content and the a The

Internet’s evolution into a worldwide societal phenomenon has much to credit to the content and services

that have taken advantage of the network’s unique architecture. Entire economies depend on cross-border

content flows. Daily innovations have the potential to disrupt entire industries. The Internet is now a critical

part of democratic processes and policy discussions. Personal relationships are created and broken online

The trend is not slowing down. According to estimates, Global Internet traffic in 2020 will be equivalent to

95 times the volume of the entire global Internet in 2005. The number of devices connected to IP networks

will be three times as high as the global population in 2020. Yet, the Internet also contains content that

policy makers, legislators, and regulators around the world want to block. From blocking foreign gambling

websites in Europe and North America to blocking political speech in China, the use of Internet content

blocking techniques to prevent access to content considered illegal under certain national laws is a

worldwide phenomenon. Public policy motivations to block Internet content are diverse, ranging from

combating intellectual property infringement, child abuse material and illegal online activities, to protecting

The detection of the data available in the images is very important. The data that the image contains

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

Page |6
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Chapter 2
PROBLEM DEFINITION AND OBJECTIVE

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 to the features and help in completing the face detection

for proving the age detection of the person in the image Over the last decade, the rate of image uploads to

the Internet has grown at a nearly exponential rate. This newfound wealth of data has empowered computer

scientists to tackle problems in computer vision that were previously either irrelevant or intractable.

Consequently, we have witnessed the dawn of highly accurate and efficient facial detection frameworks that

leverage convolutional neural networks under the hood. Applications for these systems include everything

from suggesting who to “tag” in Facebook photos to pedestrian detection in self- driving cars. However the

next major step to take building off of this work is to ask not only how many faces are in a picture and where

they are, but also what characteristics do those faces have. The goal of this project do exactly that by

attempting to classify the age and gender of the faces in an image. Applications for this technology have a

broad scope and the potential to make a large impact. For example, many languages have distinct words to

be used when addressing a male versus a female or an elder versus a youth. Therefore, automated translation

services and other forms of speech generation can factor in gender and age classification of subjects to

improve their performance. Also ,having an idea about the age and gender of a subject makes the task of

recognizing that subject significantly easier. This

Page |7
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

websites like Facebook. The objective of this paper is neither to assess such motivations nor to qualify

whether a certain type of blocking is good or bad from an ethical, legal, economic, political or social

perspective. Instead, we will provide a technical assessment of the benefits and drawbacks of the most

common blocking techniques used to prevent access to content deemed illegal. The aim is to help readers

understand what each technique can, and cannot, block, along with the side effects, pitfalls, trade-offs, and

associated costs .Our conclusion is that the use of Internet blocking to address illegal content is generally

inefficient, often ineffective, and prone to cause unintended collateral damages to Internet users,

summarized further in the table below Literature Review. Referring to the report [1], Poland is one of the

leading countries where internet advertisements are often blocked by users using ad blocking software. The

phenomenon is not conducive to the development of the electronic business of companies that earn money

by displaying ads. The perspective of charging instead of advertising is not beneficial either for publishers

or for content recipients. Therefore, changes in the sustainable advertising market are necessary. However,

to enter them, it is necessary to thoroughly understand the recipients of the ads, their reaction to the ads and

expectations related to the ads. The most popular forms of online advertisements are banners, which

are referred to as “conventional advertisements”. They consist in placing graphic elements of various

forms and sizes on a website. The elements redirect the interested party to the advertiser's website. Such a

form of advertising allows a company to establish its online presence and enhance its image. Static banners

are commonly considered the least invasive ads. In contrast, the most aggressive form of advertising is sound

ads. An interesting type of advertising is the so-called native ads. Native advertising, cleverly woven into

some content, e.g. an article, is usually not treated by the recipient as an invasive advertisement. It has

therefore become popular among publishers. In order to block advertising on a website, users use

software call.

Page |8
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

of contributors, who use a variety of undocumented heuristics to determine what rules should be included

[11]. Other comparisons of online ad blocking lists revealed that blacklists can be reactive or proactive

in tackling the online ad and tracking services There are some enterprise network environment solutions

for DNS Adblock, which block advertisement requests on the internet connection level. They reduce

traffic and bandwidth utilization [13]. Ad blocking is often analyzed in terms of digital advertising, privacy

preserving, ad effect measurement and ad fraud [14]. Recently an Adgraph tool was presented which creates

a graph representation of the HTML structure, network requests, and JavaScript behavior of a webpage,

and, using this unique representation, trains a classifier for identifying advertising and tracking resources

[15]. Nowadays, internet users are more and more resistant to different forms of advertisements. Users have

become better prepared for receiving internet promotion of products and services. They seem to be better

focused on reviewing certain criteria of advertisements, i.e. location, contrast and size of the advertising

unit [16], than during the initial period of internet development. The advertisements’ forms are still

very popular but no longer as efficient used to be. In the users’ opinion, they express strong negative

sentiment on ads and moderate positive sentiment if it is possible for them to subscribe to an ad-free fee-

financed website [17]. Adoption of ad blockers by users is positively influenced by the level of knowledge

of their advantageous features Website Owners Preventing Ad Blocking Some descriptions of cases are

characterized on the internet by the users [19] in the context of website publishers preventing them from

blocking advertisements [20]. One prevention option is based on legal regulations and the right to maintain

the integrity of the website publisher’s work, or the right not to allow the ad blocking, because it may be

viewed as a change to the content and form of the website publisher’s work [21], [22]. Another option is

based on the intention of the publisher to establish cooperation with creators of software for blocking

advertisements [23] and to create advertisements which users are unable to filter with blocking software

[24]. Nowadays, programming techniques allow detecting mechanisms for blocking advertisements very

Page |9
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

exploring a complex topic [17], [18]. The topic of advertisement blocking should be considered from

multiple perspectives. The last PageFair-2017 report showed [1], that the main reasons for blocking

advertisements using ad blocking software were [19]: security, interruption, speed, too many ads, privacy,

advertising in continuous browsing of online content. There is more than one reason for blocking the

advertisements. Over 70% of respondents chose more than one reason as “the most important one” in

connection with their use of ad blocking software. Users do not mind advertisements as such, but they are

bothered rather by their aggressive form, such as a sudden sound or an advertisement suddenly covering the

browsed content and, additionally, the advertisement not allowing itself to be skipped or closed. The Page

Fair report’s index of blocking

ads for Poland in 2017 [1] amounted to 33%, while in 2015 it maintained the level of 34.9%. Another study

published in 2018 showed that 61.2% of Polish internet users are using ad blocking software [41].

Apparently, the reason for this phenomenon is the overload with advertisements. Greece, Poland, Germany

and Sweden are the countries in Europe where the most ads are blocked. According to the PageFair report

from 2017, Internet users blocked ads mainly due to: (1) security, (2) interruption by ads of various types

of content, (3) slow loading of web pages and inconveniences connected with it, e.g. nuisance and

excess ads that pop up at any time, from all directions, and still run sounds, etc. Our own research was

slightly different from the Page Fair survey [42]. According to our research, users are most disturbed

by: (1) advertising interruption of various web content, (2) a large number of ads (excess advertising) and

(3) slow page loading (speed Research Methodology of Adblock Software Usage FactorsThe main research

goal was focused on describing the ad blocking software usage factors for future web advertisement

strategy development. Research study was started after preliminary research on types of online

advertisement, and the diversity of ad blocking reasons and methods were characterized [3]. The new

classification of ad blocking reasons was developed, based on previous studies and literature reviews [3],

Page |10
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

[42]. The study of classification included the technical and social reasons for ad blocking as well (Figure

1). After the classification study, the following research questions were proposed to fulfill the research

goals: RQ1: Who is using the adblock software? RQ2: What are the main reasons for blocking the online

advertisements? RQ3: What types of online advertisements are acceptable for users? RQ4: Who is using

usage of ad blocking software? In order to recognize the activity of users of computer programs blocking

the content of online advertisements, an electronic questionnaire was developed to discover the main

factors for adblocking software usage. The study was conducted in 2017. The target group of the study

included people of different ages. In total, responses were collected from 774 respondents. The results are

characterized by the following factors in five. Technical and Social Reasons for Blocking Online

Advertisements The respondents had learned about the possibility of blocking advertisements in the

following ways: from friends (31.5%), from the internet (22.7%), from the browser settings (8.7%). At the

same time, 35.4% of respondents did not remember where they found out about ad blocking software.

The results presented further concern only people who used ad blocking software, i.e. 596 respondents

(out of 774). Each of the 596 respondents (100%) had ad blocking software installed on a desktop computer,

99 respondents (16.6%) ‒ on a smartphone, and 54 respondents (9.1%) ‒ on a tablet. Most respondents, i.e.

400 (67.1%) used the default ad blocking settings, and the rest of the respondents had configured the

software according to their needs. Out of 596 respondents, only 77 (12.9%) did not exclude

this program. Most respondents, i.e. 519 (87.1%) actively disabled the adblocks to gain access to blocked

content. There are at least two reasons why users of content blockers temporarily disable them: • firstly,

there is a need to temporarily use content that is unavailable or invisible due to ad blocking activity (301

respondents, 58%); • secondly, it is about permanently turning off an adblock for a selected website or

adding that website to the list of exceptions (218 people, 42%). If we divide the group of respondents by

gender, it turns out that all males know what an adblock is and either use it (78.3%) or not. Among females,

Page |11
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

who receives the advertisements. At this stage, this solution seems to be the best, which limits the use of ad

block extensions that block advertisements and at the same time emphasizes the benefits of partners'

cooperation for a sustainable advertising market to gain access to content, and in the group of people aged

25–34 – 88.5% of users disabled the ad block software. Both groups of respondents react similarly to

blocking advertisements. After the collision with an(constituting the majority) decide to disable the blockade

to access the content to gain access to content, and in the group of people aged 25–34 – 88.5% of users

disabled the ad block software. Both groups of respondents react similarly to blocking advertisements. After

the collision with an adblock-wall, in the group of 18–24% up to 35.4% of the respondents would leave the

website, and in the group of 25–34 years, as many as 49.7% of the respondents would leave the website.

Other respondents (constituting the majority) decide to disable the blockade to access the content example,

if a picture contains many people studying together, Facebook might be able to caption the scene with “study

session.” However if it can also detect that the people are all men in their early 20s and that some are wearing

shirts with the same letters, it may predict “College students in a fraternity studying.” Age and gender

classification is an inherently challenging problem though, more so than many other tasks in computer

vision. The main reason for this discrepancy in difficulty lies in the nature of the data that is needed to train

these types of systems. While general object classification tasks can often have access to hundreds of

thousands, or even millions, of images for training, datasets with age and/or gender labels are considerably

smaller in size, typically numbering in the thousands or, at best, tens of thousands. The reason for this is that

in order to have labels for such images we need access to the personal information of the subjects in the

images. Namely we would need their date of birth and gender, and particularly the date of birth is a rarely

released piece of information .Sidebar : filtering , Blocking When describing Internet filtering, terms such

as “filtering,” “blocking,” “shut down,” and

“censorship” all come up (along with several others). From the point of view of the user, the

Page |12
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

term chosen is less important than the effect: some part of the Internet is inaccessible. For policy makers

and digital activists, choosing a particular term is usually more driven by semantic overtones than technical

correctness. The word “censorship” carries a strong negative connotation, while “filtering” seems a more

gentle and harmless operation, like removing unwanted seeds from a glass of orange juice. We have chosen

to use “blocking” as a simple and straightforward term throughout this paper the study involves a proper

recording of the variations of people on the basis of their changes according to age, gender, identity and

other features. The brain activation tasks related to face matching are performed and tested outside the

scanner. There was a same result in face processing in older as well as young adults. The performance results

high in both the cases having same facial viewpoints. The aging of the elders is not based on any one factor.

It is combination of various factors that result in accountancy of such results. The results need to be kept a

frequently as a part of (or together with) a facial recognition structure. Some current digital cameras use face

detection for autofocus. Face detection is also beneficial for choosing areas of interest webcam can be

integrated into a television and detect any face that walks by. The system then computes the race, gender,

and age range of the face. Once the data is composed, a series of announcements can be played that is

specific toward the detected race/gender/age. This paper shows prototype or partial application of this type

of work. Face detection is also being studied in the area of energy conservation. Procedure for face

recognition based on information theory method of coding and decoding the face image is discussed in

[Sarala A. Dabhade & Mrunal S. Bewoor, 2012][4]. Proposed methodology is connection of two stages –

Face detection using Haar Based Cascade classifier and recognition using Principle Component analysis.

Various face detection and recognition methods have been evaluated [Faizan Ahmad et al., 2013] and also

solution for image detection and recognition is proposed as an initial step for video surveillance.

Implementation of face recognition using principal component analysis using 4 distance classifiers is

Page |13
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

based on speech signals was recently proposed that utilized a fast Fourier transform (FFT) with a 30 ms

Hamming window and a 20 ms overlap for signal analysis. Each 10 ms spectrum was further filtered to

adhere to the Mel Scale, yielding a vector of 20 spectral coefficients. The mean and variance of the

Multilayer Feature Switch Card (MFSC) vectors were determined in each window, and 40 coefficients were

concatenated to generate a feature vector. The feature vector is then normalized to ensure that the classifier

captures the relationship between the frequency of the spectrum rather than the individual, Smilarly, Martin

A. F. et al. [19,21] provided one of the most innovative revelations in the field of voice or speech signal-

based gender and age identification. The study suggested multiple linguistic detections for multiple speakers.

Rather than focusing on a single speaker, multi-speaker analysis was used in this study to address the

practical environment with higher efficiency and to obtain a reasonable recognition rate. Khan et al.

suggested a fuzzy logic-based gender classification structure that was trained using several metrics, such as

the power amplitude, total harmonic distortion, and power spectrum. Although the study yielded positive

results, as the rule base grows when utilizing the fuzzy logic, the scheme becomes more complex and lacks

the added benefit of learning owing to these issues; the proposed method failed to achieve a high accuracy

Nowadays, researchers have used deep learning approaches to recognize emotions in speech and human

action in the frame of sequence In this advance era, researchers have proposed hybrid methods [28,29] and

systems to address the limitations of the age and gender recognition domain, which appear to offer a potential

answer to the problem; however, they also significantly increase the system complexity, and the problem of

noise has not been considered. Furthermore, owing to higher complexity, these systems require more

computational time [23,30], which is addressed in this study. Furthermore, Prabha et al. [31] created a

system for classifying gender based on energy; the signal transformation was performed using FFT, and the

system secured a 93.5% efficiency. Several methods use machine learning techniques for classification

Page |14
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Chapter 3
PROBLEM DEFINITION AND OBJECTIVE
PROBLEM FORMULATION

Google checks the pages that it indexes for malicious scripts or downloads, content violations, policy

violations, and many other quality and legal issues that can affect users. When Google detects content that

should be blocked, it can take the following actions If Google suspects your site of hosting dangerous or

spammy downloads, engaging in practices that are bad or dangerous to the user, or of being hacked, you

will see a warning either in Google Search results or in your brow servector systems) Google Search

results might show labels such as "This site may harm your computer" or "This site may be hacked" next

to your site Your browser might display and when you try to open your page, either from a link in Google

Search results or anywhere .OBJECTIVE The objective of python website blocker is to block some certain

website we can distract the user during the specified amount of time In this I will block the access to the list

of some particular website during the working hours so that the user can only access those website during

the free only The working time in this python application is considered from as us wish to time period

except that time will be considered as free time placement according to a range of gender and age. But this

again will not be much effective and feasible when it comes to quickly locating exactly matching products

based on the age and gender of the customer. Thus, this project provides an innovative solution to this

problem. The solution is to have a model that will instantly classify an individual’s age and gender based on

the real-time image of the individual and will direct him by specifying the section of the mall which contains

the products based on his age and gender. A camera at the entrance of the mall may feed the model with the

real-time image. Multiple such units may be used in case of busy malls. The other problem that is being

Page|15
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

addressed is related to security. The model is trained to detect human movements of any form. A camera

will continuously monitor the secured area and the model maybe activated in the absence of the user. The

model will detect any unauthorized human movement in the secured area based on the real-time footage

provided by the camera and instantly alert the user on the security infringement.3.3.GoalsMy first objective

in this project was to determine if the proposed network architecture(seeSection3.1)was indeed optimal.

Although the authors of [12] claimed that any deeper network would suffer from over-fitting, I wanted to

verify this for myself. To this end I experimented with adding additional convolution layers, removing fully

connected layers (in the style of [19]), and modifying the parameters used for dropout as well as LRN. The

primary goal, however, was to experiment with a new higher-level approach for composing these classifiers

to improve performance. The observation I made early on was that gender classification is an inherently

easier task than age classification, both due to the fewer number of classes to distinguish between and the

more marked differences that exist between genders than between many age groups. This then led me to the

conclusion that while it is reasonable to assume one should be able to ascertain someones gender apart from

knowing their age, or vice versa, there is also some plausibility of using one of these attributes to better

inform the prediction of the other. For example, the amount of hair on a man’s head can often be a useful

indicator of age, but the same is not true for women.3.4 Motivations for Blocking Content In this paper, we

focus on blocking based on public policy considerations and its effects on the Internet and Internet users

see side-bar for other motivations for content blocking speaker age, gender, speaking style, emotional

content, etc. One of the reasons Blocking based on public policy considerations is used by national

authorities to restrict access to information (or related services) that is either illegal in a particular

jurisdiction, is considered a threat to public order, or is objectionable for a particular In this paper, we focus

on blocking based on public policy considerations, but there are two other common reasons that network

blocking is put into place. The first is preventing or responding to network security threats.

Page|16
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Chapter-4

DESIGN AND IMPLEMENTATION

Design Each technique has both technical and policy limitations and consequences that need to be

considered when any type of content blocking is being proposed. The goal of this paper is to provide a

common way to evaluate their efficacy and side effects This paper will assess the following types of content

blocking IP and Protocol-based blocking Deep Packet Inspection-based blocking URL-based blocking

Platform-based blocking (especially search engines)DNS-based blocking We chose these five types of

blocking because they target the elements of a typical end-user cycle of finding and retrieving information,

including the use of a search engine and viewing information with a web browser or similar tool. This cycle

is very familiar to policy makers, themselves Internet users, and these are the operations that most blocking

based on public policy considerations tries to disrupt tIn the diagram to the right, we show the steps that a

typical Internet user might take to find information, as well as the kinds of blocks that have been used to

disrupt this cycle when blocking based on public policy considerations is implemented. In our diagram, an

Internet user searches for some type of content using a search engine (step 1), a common starting point. The

search engine returns a set of results (step 2), and the user selects one and clicks on the result (step 3). One

type of blocking, Platform-based Blocking, is used to disrupt this part of the cycle by blocking some results

coming back from the search engineThe user’s computer tries to find the server hosting the data in the

Internet’s DNS . A second type of blocking, DNS-based Blocking, is used to disrupt this part of the

cycleThen, the user’s web browser tries to connect to the server (step 6). This part of the cycle can be blocked

using three other types of blocking: IP and Protocol-based Blocking, URL-based blocking, and Deep Packet

Inspection-based blocking

Page|17
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Page|18
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Of course, the Internet is much more than search and web browsers, and many of the techniques discussed
below are effective at blocking more than web pages. For example, use of VPN services to encrypt and hide
traffic can often be blocked using a combination of Deep Packet Inspection-based blocking and IP/Protocol-
based blocking These types of blocks may be applied very specifically (such as a particular document on a
particular web site) or very generically (such as “material on an issue” or “Voice over IP services” Note that
in the case of blocking based on public policy considerations, the majority of measures are being applied at
the first two levels (national, carrier, and ISP levels)The diagram below summarizes some of the main
locations where blocking can occur, and which types of blocking can occur at each point. A SVM model is
a representation of the examples as points in space, mapped so that the examples of the different categories
are isolated by a clear gap that is as wide as would be prudent. In addition to performing linear
characterization, SVMs can effectively play out a non-linear grouping utilizing what is known as the kernel
trick, implicitly mapping their inputs into high-dimensional element spaces. SVM is a classifier derived
from statistical learning theory. It is a critical and an active field of each Machine Learning research. The
SVM classifier has some primary features. By utilizing the kernel trick, information is mapped onto a high-
dimensional element space without a great part of the computational endeavors. SVM requires that every
data instance is represented to as a vector of real numbers. Consequently, if there are categorical attributes,
one needs to change over them into numeric information. Scaling before applying SVM is critical. The
principle favorable position of scaling is to maintain a strategic distance from attributes in more prominent
numeric ranges commanding those in smaller numeric ranges. Another preferred standpoint is to

Page|19
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Endpoint Content Blocking This paper focuses on Internet content blocking based on public policy

considerations Yet, it is important to note that one of the most effective ways to block undesired content is

through the use of software installed on the user’s device, commonly called the “endpoint” because it is the

last point of the connection between the user and the Internet. Most computer users make use of endpoint

software to block malware (viruses, Trojan horses, and phishing), whether installed personally or by an

organizational IT group Endpoint content blocking software is also used by organizations to block content

for other reasons. For example, libraries often install this type of software on public computers to block the

viewing of pornography by patrons, and parents may use it to block unwanted content from their children

Endpoint content blocking may use many of the techniques described in this paper, including content

scanning, URL categorization, IP address blocking, and DNS interception. Generally, the blocking and

analysis occurs on the actual endpoint. However, vendors of this software are increasing also using cloud-

based tools including content scanning and DNS-based blocking, in cooperation with a small amount of

endpoint software. In these newer solutions, some or all of the Internet content may pass through a cloud-

based service. The advantage of moving the decision-making to the cloud is that endpoints do not have to

be constantly updated, and the performance impact of evaluating content is moved from the user’s computer

or smart phone to an easily scaled cloud of computers. When traffic is routed through a third party, though,

this also creates privacy issues by making the content available to the third party and, if poorly implemented,

security issues arise well Content Blocking Types Evaluated The five common content blocking types are

distinct in what they block and how they operate Below, the content blocking techniques are discussed in

greater detail and are evaluated against four specific criteria Which sets of users and Internet services are

affected by this technique? What sets are unaffected? How specific is the technique in preventing access to

particular content? How much collateral damage (unintended blockage) is created by this blocking

Page|20
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

What nontechnical issues, such as impact on trust and fundamental rights, are raised in using this technique?

IP and Protocol-Based Blocking IP-based blocking places barriers in the network, such as firewalls, that

block all traffic to a set of IP addresses. Protocol-based blocking uses other low-level network identifiers,

such as a TCP/ IP port number that can identify a particular application on a server or a type of application

protocol. These simplest approaches to blocking content don’t actually directly block content they block

traffic to known IP addresses or TCP/IP ports or protocols associated with some content or an application.

IP and protocol-based blocking may also be done by software on user’s computers, typically for network

security purposes For example, if the goal was to block all content hosted in the mythical country of Elbonia,

IP blocking could be used if the set of all IP addresses hosting content in Elbonia were known. Similarly, if

the goal was to block all VPN services (which are used to encrypt traffic and hide both the destination and
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
Page|21

(such as peer-to-peer file sharing or Voice over IP [VoIP] traffic) and data file types (such as multimedia

files).

Several of the techniques discussed in this paper, including Deep Packet Inspection (DPI)-based blocking

and URL-based blocking, have a very real limitation: they must be able to see the traffic being evaluated.

Web servers that offer encryption or users who add encryption to their communications (typically through

application-specific encryption technology, such as TLS/SSL) cannot be reliably blocked by in-the-network

 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
Page|22

vendors have developed some ways of identifying some types of traffic through inference and analysis, these

techniques often are simply guessing at what type of traffic they are seeing. In recent research, 49% of US

web traffic (by volume) was encrypted in February, 2016. This traffic would be effectively invisible to URL-

based blocking and DPI tools that look at content, because the only visible information

would be the domain name of the server hosting the information. To compensate for this “going dark,” some

network blocking uses active devices (called proxies) that intercept and decrypt the traffic between the user

end encryption model of TLS/SSL. When proxies are used, these cause significant security and privacy

concerns. By breaking the TLS/SSL model, the blocking party gains access to all encrypted data and can

inadvertently enable third-parties to do the same. The proxy could also change the content. If the blocking

party has control over the user’s system (for example, a corporate-managed device would be highly

controlled), the proxy may be very transparent. Generally, however, the presence of a proxy would be

obvious to the end user, at least for encrypted (TLS/SSL) traffic (e.g. the user may get an alert that the

certificate is not from a trusted authority). In addition, new industry and IETF standards (such as HTTP

Strict Transport Security [RFC6797], HTTP Public Key Pinning [RFC 7469], and DANE [RFC 6698]) and

new security features in modern Internet browsers make it more difficult to proxy (and decrypt) TLS/SSL

traffic without the knowledge and cooperation of the end user. Proxies installed for content blocking reasons

may also introduce performance bottlenecks into the flow of network traffic, making services slow or

unreliable. DPI blocking is very commonly used in enterprises for data leakage protection systems, anti-

spam and anti-malware (anti-virus) products, and traffic prioritization (such as boosting the priority of

enterprise videoconferencing) network management. However, it can also be used for more policy-based

blocking purposes. For example, use of VoIP services not provided by the national telecommunications

Internet. When the traffic is encrypted, as it often is, DPI blocking systems may no longer be effective.
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
Page|23

URL

blocking requires the blocking party (such as the user’s ISP) to have the ability to intercept and control

traffic between the end-user and the Internet. URL blocking is usually expensive, because the filtering device

generally has to be in-line between the user and the Internet, and thus requires a high level of resources to

give acceptable performance.URL blocking is generally considered to be very effective at identifying

content that may be on different servers or services because the URL doesn’t change even if the server
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
Page|24

very complicated or change frequently. This can happen because an information publisher has deliberately

high-level URLs, such as a particular web page, but is not as effective when deep links (such as individual

bits of content within a web page) are considered. Depending on how the user navigated to the particular

content, URL blocking may or may not be able to block all access—if the user has a “deep link” not covered

by the URL filter, the content will be allowed. For example, the Playboy web site includes both playboy.com

URLs, but also embedded content using the “playboy.tv” domain name. A URL filter that didn’t also include

“playboy.tv” URLs would not block the video content. All types of URL blocking are highly dependent on

the quality of the filter, and a poorly designed or overly broad filter may block unintended traffic or have

other negative effects on the user experience, such as affecting the loading or formatting of web pages when

some component is being blocked. As with Deep Packet Inspection types of blocking, URL blocking

requires some type of proxy to see the full URL when traffic is encrypted with HTTPS (TLS/SSL). See

sidebar “Encryption, Proxies, and Blocking Challenges”, page 15, for more information on the effects on

end-user privacy. For encrypted traffic, URL blocking can only see the IP address of the server, and not the

full URL, resulting in a much higher level of unintended blocking. Because proxies are expensive and

intrusive to the user experience, URL blocking does not work well as a tool for policy based blocking

Platform-Based Blocking (Especially Search Engines)In some cases, national authorities will work with

major information service providers to block information within their geographic region without blocking

the entire platform. The most common examples of platform filtering are through the major search engine

providers and social media platforms. Recently, it has also been reported that mobile application stores (such

as the Apple Store and Google Play) are working with national authorities to block downloads of specific

applications in their country

 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
page|25

While search engine blocking is the most common type of platform blocking, other platforms with

enormous users communities are often considered for this technique. Common examples of these types of

platforms include Facebook (which has over 1.5 billion active users each month) and YouTube (with over

a billion unique users). Attempts to use network-based or URL-based techniques to block individual content
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
Page|26

elements, such as a particular news article, are very difficult. Because they don’t want to be seen as blocking

all of Facebook (for example), national authorities have proposed working with major platform providers to

filter out specific types of content they deem illegal. Very little is known about the effectiveness, scope, or

as malware and pornographic material) and provide customized content feeds to their users, information on

national-specific blockages is not available. Platform-based blocking is a technique that requires the

assistance of the platform owner, such as a search engine operator like Google or Microsoft. In this

technique, queries from a particular set of Internet users to a search engine will receive a different set of

results from the rest of the Internet— filtering out pointers to content that are, in some way, objectionable.

In some cases the definition of what is to be blocked is based on local regulation and government

requirements, but it may also be due to concerns by the search engine operator. For example, a search engine

may block pointers to malware or content considered inappropriate according to its own terms of service.

Because search engine blocking requires the cooperation of the search engine provider, this limits its use to

two very specific scenarios: country-level rules (blocking content based on country-specific or region

specific rules) and age-based rules (blocking material inappropriate for young people).Search engine

blocking only affects users who choose a particular search engine, and only when the users are identified as

being from a particular set with filter rules. In age-based blocking, such as Safe Search[6] (offered by major

search engines and content providers), an explicit opt-in is required. Since search engine blocking only

filters out pointers to content, and not actual content, it is an extremely ineffective technique, and can have

the unintended consequence of drawing increased attention to the blocked content. The presence of multiple

search engines, as well as alternative methods of finding content, make this type of blocking very difficult

to enforce. Although search engine blocking seems like it does very little towards blocking content, the

technique is extremely popular at the national level, and governments around the world are known to demand
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
page|27

its use to two very specific scenarios: country-level rules (blocking content based on country-specific or

region specific rules) and age-based rules (blocking material inappropriate for young people).Search engine

blocking only affects users who choose a particular search engine, and only when the users are identified as

being from a particular set with filter rules. In age-based blocking, such as Safe Search (offered by major

search engines and content providers), an explicit opt-in is required. Since search engine blocking only

filters out pointers to content, and not actual content, it is an extremely ineffective technique, and can have

the unintended consequence of drawing increased attention to the blocked content. The presence of multiple

search engines, as well as alternative methods of finding content, make this type of blocking very difficult

to enforce. Although search engine blocking seems like it does very little towards blocking content, the

technique is extremely popular at the national level, and governments around the world are known to demand

that major search engines implement filters according to their regulations, such as infringement of copyright

or particular types of speech prohibited by national law. For example, Google reported in 2015 that it had

received 8,398 requests from 74 national courts to remove 36,834 results from its search results. Copyright

infringement requests made by individuals are also very popular: in June 2016, Google reported that 6,937

copyright owners had requested over 86 million search results to be removed from Google results during

that month .Search engine blocking is also used by individuals as part of the so-called “right to be forgotten,”

with over a million URLs globally requested to be blocked in the last two years (May 2014 to June

2016).DNS-Based Content Blocking DNS-based content blocking avoids one of the problems with other

techniques: the cost and performance impact of filtering all network traffic. Instead, DNS-based content

blocking focuses on examining and controlling DNS queries. With DNS-based content blocking, a

specialized DNS resolver (see Sidebar: DNS Overview) has two functions: in addition to performing DNS

lookups, the resolver checks names against a block list. When a user’s computer tries to use a blocked name,
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
Page|28

you learn Git, try to clear your mind of the things you may know about other VCSs, such as CVS, Subversion

or Perforce — doing so will help you avoid subtle confusion when using the tool. Even though Git’s user

interface is fairly similar to these other VCSs, Git stores and thinks about information in a very different
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Chapter 5
SYSTEM REQUIREMENT ANALYSIS

Git doesn’t think of or store its data this way. Instead, Git thinks of its data more like a series of snapshots

of a miniature filesystem. With Git, every time you commit, or save the state of your project, Git basically

takes a picture of what all your files look like at that moment and stores a reference to that snapshot. To be

efficient, if files have not changed, Git doesn’t store the file again, just a link to the previous identical file it

has already stored. Git thinks about its data more like a stream of snapshots. Nearly Every Operation Is

Local Most operations in Git need only local files and resources to operate — generally no information is

needed from another computer on your network. If you’re used to a CVCS where most operations have that

network latency overhead, this aspect of Git will make you think that the gods of speed have blessed Git

with unworldly powers. Because you have the entire history of the project right there on your local disk,

most operations seem almost instantaneous. For example, to browse the history of the project, Git doesn’t

need to go out to the server to get the history and display it for you — it simply reads it directly from your

local database. This means you see the project history almost instantly. If you want to see the changes

introduced between the current version of a file and the file a month ago, Git can look up the file a month

ago and do a local difference calculation, instead of having to either ask a remote server to do it or pull an

older version of the file from the remote server to do it locally.This also means that there is very little you

can’t do if you’re offline or off VPN. If you get on an airplane or a train and want to do a little work, you

can commit happily (to your local copy, remember?) until you get to a network connection to upload. If you

go home and can’t get your VPN client working properly, you can still work. In many other systems, doing
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Chapter-6

Software Requirement

Requirement Analysis The study involves a proper recording of the variations of people on the basis of their

changes according to age, gender, identity and other features. The brain activation tasks related to face

matching are performed and tested outside the scanner. There was a same result in face processing in older

as well as young adults. The performance results high in both the cases having same facial viewpoints. The

aging of the elders is not based on any one factor. It is combination of various factors that result in

accountancy of such results. The results need to be kept a track on which are based on all credentials kept

in certain environments. Face detection is used in biometrics, frequently as a part of (or together with) a

facial recognition structure. Some current digital cameras use face detection for autofocus. Face detection is

also beneficial for choosing areas of interest in photo. Face detection is in advance the interest of marketers.

A webcam can be integrated into a television and detect any face that walks by. The system then computes

the race, gender, and age range of the face. Once the data is composed, a series of announcements can be

played that is specific toward the detected race/gender/age. This paper shows prototype or partial application

of this type of work. Face detection is also being studied in the area of energy conservation. Procedure for

face recognition based on information theory method of coding and decoding the face image is discussed

in [Sarala A. Dabhade & Mrunal S. Bewoor, 2012][4]. Proposed methodology is connection of two stages

– Face detection using Haar Based Cascade classifier and recognition using Principle Component analysis.

Various faced etection and recognition methods have been evaluated [Faizan Ahmad et al., 2013] and also

solution for image detection and recognition is proposed as an initial step for video surveillance.

Implementation of face recognition using principal component analysis using 4 distance classifiers.

Page|52
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

2. Lanitis et al. [5] proposed the first approach applying AAM to age estimation, which extracts craniofacial

growth and skin aging during childhood and adulthood. age-specific estimation, which is based on the

assumption that the aging process is identical for everyone appearance-specific estimation, which follows

the assumption that people who look similar tend to have similar aging processes. Zhang et al. [6] formulated

the inference of each person’s age as a warped Gaussian process (WGP) estimation problem, and developed

a multi-task extension of WGP to solve the problem. Since different individuals have different aging

processes, personalization is beneficial for age estimation.

Page|53
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

Chapter-7

RESULT AND DISCUSSION

Establish a benchmark for the task based on state-of-the-art network architectures and show that chaining

the prediction of age with that of gender can improve overall accuracy. If there had been more time, I would

have dedicated more effort towards fine-tuning the parameters and the modified architectures I experimented

with. Ad monetization relies on ads being shown and tracked (and ideally seen and interacted with). For this

to happen, not only does the ad need to appear, but the tracking impression pixel needs to fire. Ad blockers

prevent both from loading - meaning you’d make no ad revenue for that user’s session

Advertising on Google Ads is largely unaffected because Google is a member of the Acceptable Ads

program. This whitelists Google’s ads to show through some of the most popular ad blocking

programs. However, users still have the option of manually opting out. Many ad blockers let them disable .

Page|54
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

CONCLUSION AND FUTURE SCOPE

We implemented an application for age and gender prediction. The application can be used both as a Python

module imported into another project and as a standalone application with command line interface. The

application provides two different face detection method. One is based on computing histogram of oriented

gradients, the second is based on cascade of convolutional networks. Additionally the application provides

a way to filter faces which are turned away from the camera. We also enable the usage of our application in

conjunction with person detection tools by accepting a person id parameter and computing the average

result. To select the basis of our application we evaluated 3 different methods. We then trained our own

model that implements the best method using a dataset made by combining all evaluated publicly available

datasets and the dataset we manually labelled. The trained model performed better on all dataset in age

estimation. The most significant improvement was on FG-NET dataset where mean absolute error improved

from 12.85 to 6.61. In gender estimation the model performed better on most of the datasets. On UTK Face

dataset it improved from 20.45% error rate to 10.6%. Furthermore, we evaluated in detail the trained model

on the manually labelled dataset. We achieved age mean absolute error of 5.66. In some cases the mean

absolute error gets as low as 3.96. The model is also able to correctly classify all male subjects. We

demonstrated that our implementation is able to estimate age and gender in a real world situation. We also

trained one age model and one gender model using integral image as an additional color channel. This did

not improve the performance and the results on most datasets were worse that the pretrained models. It is

possible that training with integral images requires a different neural network architecture or different set of

hyper parameters.In this study, a novel CNN model with a MAM mechanism was proposed for the

classification of age, gender, and age-gender using speech spectrograms. We addressed the two main issues

Page|55
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

design of an appropriate classification model. We designed a unique MAM mechanism to efficiently manage

special and temporal salient features from the input data. Our proposed MAM uses a rectangular shape filter

as a kernel in convolution layers and consists of two separate time and frequency attention mechanisms. The

time-attention branch learns to detect the temporal cues of the input data. In contrast, the frequency attention

module extracts the most relevant features to the target by focusing on the spatial features on the frequency

axes. Both extracted features were then combined to complement one another and build more robust features

for processing in the subsequent layers. We created an FLB composed of convolution, pooling, and batch

normalization layers to extract local and global high-level features. We achieved the best recognition score

for age and gender classification using the proper combination of FLB, MAM, and the FCN with the

SoftMax classifier. We evaluated the performance and robustness of our proposed model over the Common

Voice and Korean speech recognition datasets. We trained and tested our proposed model for age, gender,

and age-gender classification problems. Additionally, we conducted experiments using the transfer learning

method to evaluate the superiority of our proposed model. Our model achieved average accuracy scores of

96%, 73%, and 76% for the classification of gender, age, and age-gender tasks, respectively, for the

Common Voice dataset. A 97% recognition rate was obtained for the gender and age classification using

the Korean speech recognition dataset. For the age-gender recognition, the highest result obtained was 90%

compared to the other results. Even though our proposed model achieved the highest classification accuracy

compared to the other models, there is still confusion between teens and twenties age groups in the Common

Voice dataset. The prediction accuracy of teens age class was lower than 50% and mostly confused with

twenties age class around 40%. It still requires conducting more research to decrease the confusion between

teens and twenties age groups. Moreover, it is difficult to directly compare the results of the Common Voice

dataset and Korean speech recognition dataset because of their age group labels difference. There is a

significant difference between the classification scores of the Common Voice and the Korean speech

Page|56
 Chandigarh Engineering College Jhanjeri
Mohali-140307
Department of Computer Science & Engineering

conservation. Procedure for face recognition based on information theory method of coding and decoding

the face image is discussed in [Sarala A. Dabhade & Mrunal S. Bewoor, 2012][4]. Proposed methodology

is connection of two stages – Face detection using Haar Based Cascade classifier and recognition using

Principle Component analysis. Various face detection and recognition methods have been evaluated [Faizan

Ahmad et al., 2013] and also solution for image detection and recognition is proposed as an initial step for

video surveillance. Implementation of face recognition using principal component analysis using 4 distance

classifiers is proposed in [Hussein Rady, 2011].

 REFERENCES

“The state of the blocked web”– PageFair 2017 Adblock Report, 2017. [2] W. Rens, “Browser forensics:

adblocker extensions,” 2017. [3] A. Strzelecki, E. Abramek, and A. Sołtysik-Piorunkiewicz, “Adblock

Usage in Web Advertisement in Poland,” Advances in Intelligent Systems and Computing, vol. 887, K.

Arai, S. Kapoor, and R. Bhatia, Eds. Springer, pp. 13–23, 2019. Available: https://doi.org/10.1007/978-3-

030-03405-4_2 [4] K. Garimella, O. Kostakis, and M. Mathioudakis, “Ad-blocking: A Study on

Performance, Privacy and Counter-measures,” Proceedings of the 2017 ACM on Web Science Conference

– WebSci ’17, pp. 259–262, 2017. Available: https://doi.org/10.1145/3091478.3091514 [5] W. Q. Ming

and R. Yazdanifard, “Native Advertising and its Effects on Online Advertising,” Glob. J. Human-Social

Sci. E Econ., vol. 14, no. 8, pp. 11–14, 2014. [6] E. L. Post and C. N. Sekharan, “Comparative Study and

Evaluation of Online Ad-Blockers,” Proceedings of the 2nd International Conference on Information

Science and Security (ICISS), pp. 1–4, 2015. Available: https://doi.org/10.1109/ICISSEC.2015.7370988

[7] M. Malloy, M. McNamara, A. Cahn, and P. Barford, “Ad Blockers: Global Prevalence and

Impact,” Proceedings of the 2016 ACM on Internet Measurement Conference – IMC ’16, pp. 119–125,

2016. Available: https://doi.org/10.1145/2987443.2987460 [8] J. Huang and W. Cheng, “Filtering

performance analysis and application study of advertising filtering tools,” Proceedings of the 12th

International Conference on Intelligent Systems and Knowledge Engineering (ISKE), pp. 1–5, 2017.

Available: https://doi.org/10.1109/ISKE.2017.8258716 [9] M. Ikram and M. A. Kaafar, “A First Look

at Mobile Ad-Blocking Apps,” Proceedings of the IEEE 16th International Symposium on Network

Computing and Applications (NCA), pp. 1–8, 2017. Available:

https://doi.org/10.1109/NCA.2017.8171376 [10] M. Alrizah, S. Zhu, X. Xing, and G. Wang, “Errors,

Misunderstandings, and Attacks,” Proceedings of the Internet Measurement Conference on – IMC

Available: https://arxiv.org/abs/1810.09160 [12] S. S. Hashmi, M. Ikram, and M. A. Kaafar, “A

Longitudinal Analysis of Online Ad-Blocking Blacklists,” Proceedings of the 44th IEEE Conference on

Local Computer Networks (LCN), pp. 1–9, 2019. Available: https://arxiv.org/abs/1906.00166 [13] M. I.

Bin Samsuddin, M. Y. Darus, S. J. Elias, A. H. M. Taib, N. Awang, and R. Din, “The evaluation of
AdBlock technique implementation for enterprise network environment,” Indones. J. Electr. Eng.

Comput. Sci., vol. 13, no. 3, pp. 1102–1109, Mar. 2019. Available:

https://doi.org/10.11591/ijeecs.v13.i3.pp1102–1109 [14] B. R. Gordon, K. Jerath, Z. Katona, S.

Narayanan, J. Shin, and K. C. Wilbur, “Inefficiencies in Digital Advertising Markets,” SSRN Electron.

J., pp. 1–48, Sep. 2019, Available: https://doi.org/10.2139/ssrn.3459199 [15] U. Iqbal, P. Snyder, S.

Zhu, B. Livshits, Z. Qian, and Z. Shafiq, “AdGraph: A Graph-Based Approach to Ad and Tracker

Blocking,” Proceedings of the IEEE Symposium on Security and Privacy, pp. 1–14, 2020. Available:

https://umariqbal.com/papers/adgraph-sp2020.pdf [16] R. J. Walls, E. D. Kilmer, N. Lageman, and P.

D. McDaniel, “Measuring the Impact and Perception of Acceptable Advertisements,”