See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/341001786

Voice Based E-Mail System using Artificial Intelligence

Article · April 2020

CITATION READS

1 1,418

2 authors:

Sumit Raj Rizwan Khan

ABES Institute of Technology ABES Institute of Technology
1 PUBLICATION   1 CITATION    79 PUBLICATIONS   190 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Machine Learning View project

Automatic Test case generaion View project

All content following this page was uploaded by Sumit Raj on 29 April 2020.

The user has requested enhancement of the downloaded file.

 International Journal of Engineering and Advanced Technology (IJEAT)
ISSN: 2249 – 8958, Volume-9 Issue-3, February, 2020

Voice Based E-Mail System using Artficial

Intelligence
Rijwan Khan, Pawan Kumar Sharma, Sumit Raj, Sushil Kr. Verma, Sparsh Katiyar

 emails and cannot read the information shared through email;

Abstract: One of the mostly used forms of communication therefore the existing systems are not easily accessible to
among the people is Email. Lot of confidential and urgent them.
information is exchanged over emails in today’s time. There are
about 253 million visually impaired people worldwide. These
To access the internet the person must be able to read what is
visually impaired people are facing a problem of communication. written on the screen so, this makes internet useless
Since, technology is growing day by day these types to visually technology for visually challenged people. There is only one
challenged people feel that they are more challenged. So authors way by which a visually challenged person can send an E-mail
proposed a Voice based Email System using AI that will make is, they have to tell the entire content of the mail to a third
email system very easily accessible to visually challenged people person so that the third person can compose the mail and send
and also help society. Accessibility is the most important feature on the behalf of the visually challenged person [10]. But this
that is considered while developing this system. Any system is approach does not take us to the solution of the problem.
called accessible only if both able and disable people can use it
easily.
Every time finding a third person is not possible for a visually
challenged person and also sometimes the content can be
Keywords: Speech recognition, Text to speech, Voice mail, personal, for mmaintaining the Integrity of the Specifications.
visually challenged people. Therefore, for helping these people and developing society
authors have come up with this idea that helps a visually
I. INTRODUCTION challenged person by providing ability to send and receive
emails throw voice commands without using any keyboard
The technologies are growing very fast day by day this and visual thing.
has made the lifestyle of people so easy as mostly all work can Artificial Intelligence for Speech Recognition: Artificial
be done in less amount of time with accuracy and efficiency. intelligence (AI) is a technology used for creating intelligent
Communication is one of those fields that have grown to next systems and machines that simulate human intelligence. Some
level with the advancement in technology and the availability artificial intelligence applications include various expert
of Internet. Technologies have made communication so easy systems, natural language processing (NLP), machine vision
that distance has become a negligible parameter in and speech recognition. To understand and analyse human
language such as English by extracting metadata from
communication [9]. When we think of communication using
keywords, emotions, relations and concepts is Natural
internet, the first thing that we come in our mind is
Language Processing [13, 14, 15].
communication via email. Email is one of the most reliable
ways for exchange of some important information and also
II. EXISTING SYSTEM
email is used worldwide, but for accessing internet a person
must be able to see. There are millions of people who are According to Email Statistics Report, 2014-2018 by a
blind or visually challenged who are not able to see the technology market research firm Palo Alto, CA, USA, there
screen; keyboards therefore they are not able to access the are a total of 4.1 billion email accounts created until 2014 to
internet [12]. In this way, they are very far away from email over 5.2 billion accounts in end of 2018 and making it one of
communication and internet world. These blind people cannot the most used form of communication[1].
use the existing email system, they cannot send, receive The research, by the Vision Loss Expert Group (VLEG),
shows that worldwide 253 million people are either blind or
visually challenged that is, around 253 million people are not
Revised Manuscript Received on February 13, 2020. aware of how to use Internet or E-mail [6].
* Correspondence Author Existing systems of today are basically applications that
Rijwan Khan*, Department of Computer Science and Engineering,
provide accessing and managing of emails benefits to its users
ABES Institute of Technology, Ghaziabad, Affiliated to AKTU Lucknow,
India. E-mail: rijwankhan786@gmail.com via web facilities [7]. Making email widely used
Pawan Kumar Sharma, Department of Applied Science, Krishna communication form. The existing systems do not support any
Engineering College, Ghaziabad, Affiliated to AKTU Lucknow, India. voice commands or audio facilities and therefore it is not
E-mail: vashistha23@gmail.com suitable for visually challenged people.
Sumit Raj, Department of Computer Science and Engineering, ABES
Institute of Technology, Ghaziabad, Affiliated to AKTU Lucknow, India. Also, there are various existing search engine which take
E-mail: sumitraj20399@gmail.com request in form of text from user and retrieve the relevant
Sushil Kr. Verma, Department of Computer Science and Engineering, documents from server and responds by displaying it in the
ABES Institute of Technology, Ghaziabad, Affiliated to AKTU Lucknow, form of text which is not possible for visually challenged
India. E-mail: sushilssm92@gmail.com
Sparsh Katiyar, Department of Computer Science and Engineering, people [8].
ABES Institute of Technology, Ghaziabad, Affiliated to AKTU Lucknow,
India. E-mail: sparshktyr@gmail.com

Published By:
Retrieval Number: C5930029320/2020©BEIESP
2277 Blue Eyes Intelligence Engineering
DOI: 10.35940/ijeat.C5930.029320 & Sciences Publication
 Voice Based E-Mail System using Artficial Intelligence

Although the existing web browsers can play audios and The main advantage of this system is that the use of a
videos, but for that also the user has to request by typing some keyboard is completely eliminated; the user will have to
text to search after that the user will be able to play the audio respond through voice only.
and videos by using Graphical User Interfaces (GUI) [11]. The application is totally voice-based allowing blind
The existing email system do not provide this facility and people to send and receive emails easily. It converts the user
are not accessible by blind people therefore the system which spoken voice into text and the text to voice and performs the
is developing now completely different from exiting system action accordingly.
[2].
IV. DESIGN OF PROPOSED SYSTEM
III. PROPOSED SYSTEM
This project is designed by dividing it into the following
The Proposed system will make the email system very three phases:
easily accessible to visually challenged people and also help 1. UI design: In this phase, the UI or the user interface of the
society. Authors proposed the system keeping one idea in project is developed. That is the designing of the web pages in
mind that it should be easily accessible for all kind of persons. which the user will use to interact. The user interface is
Any designed web application is claimed to be excellently designed using HTML5 and CSS3.
accessible, if it is used by any person, visually challenged or
2. Database design: Database is important in every project
not with efficient manner.
since it is responsible for storing of data and user credentials.
As critical the present system that prioritizes
That is, database mainly aims User authentication and storing
user-friendliness of traditional users, proposed system
all the user mails. The database design will include various
focuses on user-friendliness of all kinds’ of individuals,
tables’ creation for storing emails.
together with traditional folks and visually weakened folks
additionally as illiterate people. 3. System design: The system will consist all the modules
In this system, the pc is going to be prompting the user to such as: TTS (Text to Speech) and STT (Speech to Text)
perform specific operations to avail various services and if the module, Mail programming module (Compose, Inbox, and
user has to access the various services then he/she has to Sent Mail).
perform that operation. Firstly, the user will have to register in
application system through the registration form. The user
goes to be assisted through voice commands whereas
registering all the mandatory fields to be stuffed are going to
be scan by website; once the user would speak it would get
written automatically.
After successfully registering, the user can log in by
speaking the Username and Password when prompted by the
system, this username and password will then be converted
from speech to text and then the user will be authenticated by
verifying the credentials with the database. Users can access Fig. 2. System Design Modules
various sections like Compose, Inbox, and Sent Mail after
successful login. A. Mail Programming Module:
Email is appearing as one of the most valuable services on
the internet today. Many of the internet systems use SMTP as
a method to transfer mail from one user to another. SMTP is a
sending protocol and is used to send the mail while POP (post
office protocol) or IMAP (internet message access protocol)
are used to retrieve those emails at the receiver’s
side.

Fig. 3. Model of SMTP system

Sending Email: When an email is distributed across, it will
contain certain things like header and its own body. A
sequence of responses requested message is line up between
shopper and server in sending an email.
Fig. 1. Flow chart of User login and accessing system

Published By:
Retrieval Number: C5930029320/2020©BEIESP
2278 Blue Eyes Intelligence Engineering
DOI: 10.35940/ijeat.C5930.029320 & Sciences Publication
 International Journal of Engineering and Advanced Technology (IJEAT)
ISSN: 2249 – 8958, Volume-9 Issue-3, February, 2020

How header and body are differing from each other, header  Prosody Generation The prosody is created after the
will terminate when a null line is in it. The message body accent has been decided . The degree of openness of a TTS
contain the particular data browse by receipt. In body every system relies on speech factors like inflection modelling
data is taken after null line. (phrasing and accentuation), amplitude modelling and length
Receiving Email: In a specific time interval user agent at modelling (including the length of sound and therefore the
server aspect checks mailboxes. If in this time during any data length of pauses, that determines the length of the linguistic
receive it alert the user immediately. When user opens its unit and therefore the tempos of the speech) [3].
email he/she is availed to scan emails with some information
as in its subject line. VI. AUTHENTICATION & SECURITY
Authentication involves provide users the credentials,
V. EXPERIMENTAL SETUP
such as username and password, and verifying that user
provides correct username and password every time
A. Speech to Text by using Artificial Intelligence
whenever application needs to be accessed. Therefore, these
Audio to text can be converted by applying a simple to use credentials need to be stored in database for future
API involving powerful neural network models provided to comparisons. Authentication will be used in our system to
developers by Google Cloud Speech to Text. The API verify the users.
acknowledges a hundred and twenty languages & variants to Storing Password directly can be Risky: One simple
support world user base. You’ll alter voice command & method to store password is to simply map them by creating
management, transcribe of audio from decision centers, and a table in database. Whenever a user login request arrives,
more. It will method period streaming or pre-recorded audio, the server will be called for authentication with a payload
mistreatment Google’s machine learning technology [4]. contain a username and a password. Then, these credentials
B. Text-to-speech unification will be matched with the password stored in database. If the
match gives positive results then the user will be given
This is an automated conversion of a text into speech. This
access to the application.
technique resembles, as closely as doable, a verbalizer of the
When passwords are stored in clear text format it can be
language reading that text. Text-to-speech synthesizer (TTS)
risky and will be open for attackers. If some attacker can
is the technology that lets laptops speak to you. With some
steal the password then he can access the users account. One
good mathematical models TTS engine availed to pre
way to store password securely is to store it by transforming
processes the text and synthesizes it. The TTS engine
it into a form that cannot be converted into original
typically generates sound information in AN audio format due
password again. The process is referred as hashing.
to the output.
C. Text-To-Speech unification System Structure
Steps involved in text to speech are :-text is received by
TTS and analyze for remodeling into an accent description,
generate the prosody . Then in an exceedingly any step, it
generates the prosody. From the data currently on the market,
it will turn out speech signal. The text-to-speech’s structure
unification may be diminished into major modules:
Fig. 4. Hashing Algorithm
D. Processing of Natural Language module There are various algorithms for hashing mostly used are
It produces a written text of the scanned text, at the side of Message Digest Algorithms (MD5) and Secure Hash
prosody. Algorithms (SHA) [5].
E. Processing of Digital Signal module
VII. COMPARISON & RESULT ANALYSIS
It transforms representative data it be subject from
information science into hearable and comprehensible SR. TRADITIONAL PROGRESSIVE
speech. NO. SYSTEM SYSTEM
The operations of Processing of Natural Language module 1. Less security. High Security provided.
are as follows: 2. Depends on Keyboard. The entire structure is
Text Review: First, text is divided into tokens. The based on IVR- interactive
token-to-word transformation creates the writing kind of voice response.
token. For token “Mr” the writing kind “Mister” is made by 3. Slow processing. Faster and more efficient
enlargement, the token “11” gets the writing kind “eleven” 4. The disables cannot use The disables can use the
and “1979” is remodeled to “nineteen seventy-nine”. the normal mail system. normal mail system.
5. Blind people are not They will be able to
 Application of Pronunciation Rules: Pronunciation rules
being able to interact interact with the web
are applied after the completion of text analysis. Letters can't
with the web based email based email system.
be reworked 1:1 into phonemes as a result of the
system.
correspondence isn't forever parallel. In sure environments,
one letter will correspond to either no sound (for example, “”
in “taught”) or many phonemes (“n” in “known”).
Additionally, many letters will correspond to one sound
(“wh” in “which”).

Published By:
Retrieval Number: C5930029320/2020©BEIESP
2279 Blue Eyes Intelligence Engineering
DOI: 10.35940/ijeat.C5930.029320 & Sciences Publication
 Voice Based E-Mail System using Artficial Intelligence

On 10 scales there parameters are taken in comparison. 9. Lazar, Jonathan, et al. "What frustrates screen reader users on the web:
A study of 100 blind users." International Journal of human-computer
interaction 22.3 (2007): 247-269.
10. December, John. "Units of analysis for Internet
communication." Journal of Computer-Mediated Communication 1.4
(1996): JCMC143.
11. Pathan, Naziya, et al. "V-Mail (Voice Based E-Mail Application)."
(2019).
12. Symons, Gary Mark, and Kirk David Symons. "Digital media editing
interface using a supercursor for selecting media clips for editing." U.S.
Patent No. 8,527,879. 3 Sep. 2013.
13. Shabana, T., et al. "Voice based email system for blinds." International
Journal of Advance Foundation And Research In Science & Engineering
(IJAFRSE) Volume 1 (2015).
14. Al Smadi, Takialddin, et al. "Artificial intelligence for speech
recognition based on neural networks." Journal of Signal and
Fig. 5. Result with compariion study Information Processing 6.02 (2015): 66.
15. Senders, Joeky T., et al. "Natural and artificial intelligence in
neurosurgery: a systematic review." Neurosurgery 83.2 (2018):
VIII. FUTURE SCOPE 181-192.
16. Collobert, Ronan, et al. "Natural language processing (almost) from
There is wide future scope of this system many scratch." Journal of machine learning research 12.Aug (2011):
enhancements can be done in the system such as including 2493-2537.
different languages, including functionality of accessing the
deleted mails and spam mails. Also, this system can be AUTHORS PROFILE
enhanced such that it can also send attachments which are Dr. Rijwan Khan is B.Tech (CSE), M.Tech (CSE)
more beneficial for visually challenged people. and Ph.D in Computer Engineering. He has total 13
This system can be made available to all regional people years of teaching experience. Now working as a Head
who are not educated enough and inclusion of different of Department in ABES Institute of Technology. His
are of research is software testing, soft computing,
languages will make this system easily accessible. nature inspired algorithms. He published more than 25
Furthermore sign language system can also be integrated with research papers in different journals. He is reviewer of
the system to make the system more scalable and robust. more than 6 different Scopus and SCI indexed journals like journals of
Inderscience, Springer, and Elsevier etc. He is authors of 3 books on C
Programming, Data Structure using C, and Operating System and one book
IX. CONCLUSION chapter in springer book.
This paper proposes a system that will be beneficial for Dr. Pawan Kumar Sharma has completed his Ph.D.
society by allowing disable people also to grow along with in Reliability Theory (Operations Research). He had
society. This project makes visually challenged people able done his research work on Operations Research and
enough to be part of growing digital India by allowing them to Graph Theory where he published research papers. He
also published two research papers on Fuzzy graph &
communicate via internet and also making life of such people
Image Processing. Currently he is actively involved in
much easier. This system overcomes many drawbacks that Birla Institute of Technology & Science, Pilani
were faced by visually challenged people such as sending and (Rajsthan) as a Guest Faculty for working professionals & INMANTEC
receiving emails. Success of this project can make an impact Institutions Ghaziabad. He has more than 14 years of teaching learning &
on developers motivating them to make something useful that research experience of graduate & post graduate classes. Proven Expertise in
designing/restructuring course outline, material and assessment criteria and
can help visually challenged or blind people implementing innovative instruction methodologies to match international
standards.
REFERENCES
Sumit Raj is an undergraduate computer science &
1. G. O. Young, “Synthetic structure of industrial plastics (Book style with engineering student pursuing B.Tech at ABES
paper title and editor),” in Plastics, 2nd ed. vol. 3, J. Peters, Ed. New Institute of Technology, Ghaziabad. His Area of
York: McGraw-Hill, 1964, pp. 15–64. interest is Artificial Intelligence, Machine Learning
2. The Radicati website Email Statistics Report, 2014-2018 and Soft Computing. He done some training in
Available:http://www.radicati.com/wp/wpcontent/uploads/2014/01/Em Machine learning.
ailStatisticsReport-20142018-Executive-Summary.pdf.
3. Ingle, Pranjal, Harshada Kanade, and Arti Lanke. "Voice based e-mail
System for Blinds." International Journal of Research Studies in Sushil kr Verma is an undergraduate computer
Computer Science and Engineering (IJRSCSE) (2016): 25-30. science & engineering student pursuing B.Tech at
4. Isewon, Itunuoluwa, O. J. Oyelade, and O. O. Oladipupo. "Design and ABES Institute of Technology, Ghaziabad. His Area of
implementation of text to speech conversion for visually impaired interest is Artificial Intelligence, Machine Learning
people." International Journal of Applied Information Systems 7.2 and Soft Computing. He done some training in
(2012): 26-30. Machine learning.
5. Shakhovska, N., O. Basystiuk, and K. Shakhovska. "Development of the
speech-to-text chatbot interface based on Google API." CEUR
Workshop Proceedings. Vol. 2386. 2019. Sparsh Katiyar is an undergraduate computer science
6. https://www.juniper.net/documentation/en_US/junos/topics/concept/ip & engineering student pursuing B.Tech at ABES
sec-authentication-solutions.html. Institute of Technology, Ghaziabad. His Area of interest
7. Arlinghaus, Robert, et al. "Understanding the complexity of is Artificial Intelligence, Machine Learning and Soft
catch-and-release in recreational fishing: an integrative synthesis of Computing. He done some training in Machine
global knowledge from historical, ethical, social, and biological learning.
perspectives." Reviews in Fisheries Science 15.1-2 (2007): 75-167.
8. Cole, Ron, et al. "The challenge of spoken language systems: Research
directions for the nineties." IEEE transactions on Speech and Audio
processing 3.1 (1995): 1-21.

Published By:
Retrieval Number: C5930029320/2020©BEIESP
2280 Blue Eyes Intelligence Engineering
DOI: 10.35940/ijeat.C5930.029320
View publication stats & Sciences Publication