TEXT EXTRACTION AND VOICE SYNTHESIS
Presented by
Anusha M(4AD15CS008)
Under the Guidance of
Mr.Raghuram A S, Bhoomika H S(4AD15CS013)
Asst. Professor,Dept. Of CSE, Gagana V(4AD15CS022)
ATME College,Mysuru. Lavanya S(4AD15CS039)
Under the coordinator of Under the cordinator of
Mrs.Sunitha patel M S Mr.Anilkumar C J,
Asst Professor,Dept Of CSE, Associate professor,
ATME College,Mysuru Dept of CSE,
ATME College, Mysuru.
INDEX
Introduction
Problem Statement
Advantages
System Specifications
Methodology
Design
Implementation
snapshots
INTRODUCTION
Our project is capable to recognize the text and convert
the input into audio.
The input can be given in many formats such as text, pdf,
docx,format and image(jpg, png).
Image acquisition, recognition and speech conversion
using Optical Character Recognition (OCR) .
An Image Processing Technology used to convert the
image containing horizontal text into text documents and
the extracted text is converted into speech.
PROBLEM STATEMENT
The project is to recognize the text character of an image
and convert this text into speech signal. To achieve this,
text contained in the page is first pre-processed. The pre-
processed unit is prepared this for voice output.
REQUIREMENTS SPECIFICATION
Software Requirements
Operating system : Windows 7.
Coding Language : python 3.6
Data Base : my SQL lite
Tool : sublime text , django
Hardware Requirements
Processor : Intel i3
Speed : 2.53 Ghz
RAM : 4 GB
Hard Disk : 500 GB
Speakers
METHODOLOGY
OCR(Optical Character Recognition)
Optical character recognition, or OCR, is a method of
converting a saved image into text.
OpenCV(Open source computer vision)
It is an library using which we can develop real time
computer vision application. This library consists of
inbuilt features or functions. It mainly focuses on image
processing including features like physical object, face
and text identification and recognition.
TTS(Text to speech)
It is a type of speech synthesis application that is used to
create a spoken sound version of text in a computer
document or image.
NLP(Natural language processing)
It consists of different type of english versions. It will
match the ASCII value(extracted from text or a
document) and HMM value(from dataset for speech).
Design start
Input files
Check
extension
Image files Doc files
( jpg, png) (doc, pdf)
OpenCV
Pre processing
A
A
Recognition
Text to speech
Voice output
Stop
ALGORITHM
OPENCV
• Predefined-26 letters,0-9 numbers, special
characters.
• The predefined text is matched with input source.
• This is how the words are divided into blocks.
In binarization the pixel image is converted into
grayscale image.
Grayscale conversion
• The file size like 5mb,10mb,15mb in this the 5mb
file is processed fast compare to other.
• The pixel have its own color,angle,depth.
• If we use grayscale the size of the file is reduced.
Filter
It is used to modify or enhance the image.
Noise
Noises will be removed.
The binary input will be compared with dataset ,if
matched then the output will be stored in binary
form again.
For example:
Binary input:A-11,B-10,C-01,D-00
Dataset:A-11,B-10,C-01,D-00
Matched:A-A,B-B,C-C,D-D
NLP ALGORITHM
We use Google text to speech algorithm in text to
speech synthesizer.
The binary output will be given as the input (ASCII).
The hidden markov model (HMM) values will be
stored in the database.
Here we will match the binary output(ASCII values)
with the HMM in the database.
After matched,the digital signal processing takes
place.
And the output will be converted to analog signals.
IMPLEMENTATION
Register
• In registration page a new user can input his/her
name and he can create his own password.
• As soon as user give his/her name a unique
username will be generated by the application.
Login
• Once the user is successfully registered then user
can login into the application
• User can use his/her username generated by the
application to login to the application
• If the user fails to login he can use forgot
password option by giving the email id and the
password will be sent to the given email id.
File upload
• In this module the user can upload the file.
• As soon as the image to be uploaded a unique id
will be created and the date will be uploaded
automatically by the application.
SNAPSHOTS
