Image To Text and Speech Conversion

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Image to Text and Speech Conversion


Nishu Kumari

Abstract:- This text-to-image convertor aims to check II. OBJECTIVES


the conversion of data between the various modalities
(text, image) because of the evolution of human-machine The objective of OCR software package aims to
communication that introduced the utilization of natural organize the text and convert it to editable form. Thus,
communication modalities to humans. Such as gestures, developing pc algorithms to identify the character within the
speech, sound, and vision. In fact, one of the main text’s character is OCR. A document is 1st scanned by an
challenges of this "multimodal" learning is the learning associate degree optical scanner, which produces a picture
of a shared illustration between the distinct modalities style of it that's not editable. Optical character recognition
and the prediction of the missing knowledge ( by involves the Translation of this text image into editable
retrieval or synthesis) from one conditioned modality to character codes like code.
another. Some researchers work on the various varieties
of conversions; Text to Speech, Speech to image or Text The below diagram shows the process mechanism of
to image synthesis, and vice-versa however in this paper OCR system:
we tend to can focus on: image to audio image-to-text
synthesis.

I. INTRODUCTION

Improve the interface between man and machine in


varied Textual info is on the market in several resources like
documents, newspapers, faxes, written info, written notes,
etc. many folks merely scan the document to store the data
within the computers. Once a document is scanned with a
scanner, it's kept within the kind of pictures. however these Fig. 1: OCR Engine
pictures are not editable and it's troublesome to search out
III. LITERATURE SURVEY
what the user requires as they're going to got to undergo the
entire image, reading every line and word to work out if it's A. PREVIOUSLY PROPOSED MODEL
relevant to their need. pictures additionally take up extra
In the current world, there's a growing demand for users
space than word files on the pc. it's essential to be ready to
to convert written documents into electronic documents for
store this information in such the simplest way so it maintaining the safety of their information. Hence, the basic
becomes easier to go looking and edit the information.
Text recognition system was fictional to acknowledge and
there's a growing demand for applications that will
convert the information available on paper into laptop-
acknowledge characters from scanned documents or
processable documents, so the docu- ments will be editable
captured pictures and build them editable and easily and reusable.
accessible.
The prevailing system/previous system of Text
Character recognition is one of the foremost
recogni- tion on a grid infrastructure is simply a text
interesting areas of pattern recognition and artificial
recognition system while not grid practicality. The existing
intelligence. Optical Character Recognition extracts the system deals with the same character recognition or
relevant info and mechanically enters it into an electronic character recognition of a single language. The drawback
information service rather than the traditional way of within the early text recognition system is that they solely
manually retyping the text. Optical Character Recognition
can acknowledge and convert solely photographs of English
may be a Brobdingnagian field with a variety of various
or solely of a single language. that the older Text
applications like invoice imaging, legal trade, banking,
recognition system is Unilingual.
health care trade, etc. OCR is additionally widely used in
several alternative fields like Captcha, Institutional B. PROPOSED MODEL
repositories and digital libraries, Optical Music Recognition This planned system is that the Extraction of text from an
with no human correction or human effort, Automatic image recognizer OCR, on a grid infrastructure which will
variety plate recognition, and Handwritten Recognition. It be a personality recognition system that supports the popu-
contributes immensely to the advancement of the associate larity of the characters of multiple languages. This feature is
degree automation process and may applications. Several what we have a tendency to tend to call grid infrastructure
research works are that specialize in new techniques and that eliminates the matter of heterogeneous character
strategies that may cut back the processing time whereas recognition and supports multiple functionalities to be per-
providing higher recognition accuracy. currently, it's formed on the image. throughout this context, Grid infra-
potential to scan documents as associate degree images and structure suggests that the infrastructure supports a cluster of
to form editable and searchable for further informatics. specific sets of languages.

IJISRT22DEC1083 www.ijisrt.com 1965


Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Thus, the Extraction of text from image exploitation of E. CLASSIFICATION
for- mal logic on a grid infrastructure is multi-lingual. Image classification is that the method of categorizing
and labeling teams of pixels or vectors at intervals in a
The advantage of a planned system that overcomes the picture supported by specific rules. The categorization law is
ad- vantage of the current system is that it supports multiple de- vised mistreatment one or additional spectral or textural
functionalities sort of a piece of writing and the continua- characteristics. Two general ways of classification are 'su-
tion ion the text to the character. It to boot adds profit by pervised' and 'unsupervised'.
providing heterogeneous characterters recognition. this
technique acknowledges the characters supported by the F. POST-PROCESSING
trained info values. Post-processing is that the use of any technique or tech-
nology to boost the initial image captured by the artist. An
Working Of OCR Model old post-processing technique was airbrushing, which was
done to get rid of or soften one thing within the original
image.

IV. CONCLUSION

Nowadays, applications need several types of pictures


as sources of data for elucidation and analysis. once a
picture is reworked from one kind to a different one like
digitiz- ing, scanning, human activity, storing, etc.
degradation happens. Therefore, the output image has got to
under- take a method known as image improvement, which
con- tains a group of strategies that request to develop the
vis- ual presence of a picture. Image improvement is
enlight- ening the interpretability or awareness of data in
pictures for human listeners and providing higher input for
Fig. 2: Working of OCR differ- ent automatic image process systems. OCR image
process may be a powerful tool for kind preparation of
A. IMAGE ACQUISITION knowl- edgeable information edge and therefore the
The general aim of Image Acquisition is to remodel an combination of inaccurate data from totally different
associate degree optical image (Real World Data) into an sources. The sup- posed tesseract rules square measure a
associate degree array of numerical infor- mation that may horny result to en- hance the standard of edges the
be later manipulated on a pc, before any video or image maximum amount as po- tential.
process will begin a picture should be captured by the
camera and reborn into a manageable entity. ACKNOWLEDGMENT
B. IMAGE PREPROCESSING The authors wish to thank Er. Amandeep Kaur and Er.
Image preprocessing is that the steps taken to format Mariam Khan for their guidance.
pictures before they're utilized by model coaching and
abstract thought. This includes, however, isn't re- stricted to, REFERENCES
resizing, orienting, and color corrections.
[1.] L. Neumann and J. Matas. A method for text
C. SEGMENTATION localization and recognition in real-world images. In
Image segmentation involves changing a picture into a ACCV, pages 770–783, 2010.
set of regions of pixels that area unit delineated by a mask or [2.] Reza Sarshogh and Keegan Hines, ”Computer Vision
a labeled image. By dividing a picture into segments, you'll Methods for Extracting Text from Images”, Capital
be able to method solely the vital segments of the image One Tech.
rather than process the com- plete image. [3.] T. Som, Sumit Saha,"Han dwritten Character
Recognition Using Fuzzy Membership Function",
D. FEATURE EXTRACTION International Journal of Emerg- ing Technologies in
In computer vision and image process, a feature could be Sciences and Engineering, Volume 5, De- cember
a piece of data concerning the content of an image; typically 2011. L. A. Zadeh. Fuzzy sets, Information Control 8
about whether a certain region of the image has certain (1965) 338–353.
properties. Features could also be specific structures within [4.] Gur, Eran, and ZeevZelavsky, “Retrieval of Rashi
the image like points, edges, or ob- jects. Semi-Cursive Handwriting via Fuzzy Logic,” IEEE
International Conference on Frontiers in Handwriting
Feature extraction refers to the method of reworking Recognition (ICFHR), 2012
raw data raw information to numerical options which will be [5.] Thomas Natsvhlager, “Optical Character
processed whereas conserving the data within the original Recognition”, A Tuto-rial for the Course
data set. It yields higher results than ap- plying machine Computational Intelligence.
learning to the data. [6.] Andrei Polzounov,Arts iomAblavatski , Sergio
Escalera, Shijian Lu, JianfeiCai “Wordfence: Text

IJISRT22DEC1083 www.ijisrt.com 1966


Volume 7, Issue 12, December – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Detection In Natural Images With Border
Awareness”
[7.] D. Trier ,A.K.Jain ,TTaxt , Feature Extraction
Method for Charac- ter Recognition-A Survey” ,
Pattern Recognition
[8.] https://en.wikipedia.org/wiki/Optical_character_recog
nition
[9.] Image to Text Converter
Website https://www.prepostseo.com/image-to-text

IJISRT22DEC1083 www.ijisrt.com 1967

You might also like