You are on page 1of 1

PRoJECT PRESENTATION PRoJECT PRESENTATION PRoJECT PRESENTATION PRoJECT PRESENTATION

Objective Objective
Preprocessing is a stage in typical OCR
system which focuses on enhancing the
acquired image. The accuracy of OCR
systems in terms of recognition heavily
depends upon preprocessing stage. The
preprocessing stage consists of
removal noise, binarization and
character segmentation.
ac!ground ac!ground
"pigraphy is the study of inscriptions,
and more specifically, the deciphering
of ancient inscriptions on roc!s, pillars
and other writing material. #t is one of
the most fascinating and instructive
studies and deals with the art of writing
and provides us with an instrument for
conservation and transmission of
historical traditions from generation to
generation.
The automatic processing of degraded
historical documents is a challenge in
document image analysis field which is
confronted with many difficulties due to
the storage condition and the
comple$ity of their content. %or
historical degraded and poor quality
documents, enhancement is not an easy
tas!. The main interest of enhancement
step of historical documents is to
remove unwanted information appear in
the bac!ground and highlight the
foreground.
This pro&ect addresses the
preprocessing techniques for
handwritten 'annada documents and
helps us in integrating the new
technological concepts with underlying
cultural features of the scripts, thereby
minimizing the divide between man and
the history.
Fig-Segmentation results for
connected component and
disconnected component.
Results ( Output Results ( Output
. The preprocessing stage is mainly
designed by ta!ing into account the
images of ancient inscriptions and
segmentation is performed on present
'annada handwritten te$t documents

The project preprocesses ancient
Kannada documents along with
enhancing the quality of ancient scripts
which are degraded and non readable
due to various factors by using suitable
techniques.
Further the project also helps to
remove noise in the document images to
the maximum extent. The characters
that are segmented belong to the present
Kannada character set.
Conclusion Conclusion
)pplications of the Pro&ect )pplications of the Pro&ect
Popular applications of pattern
recognition and many systems have
been proposed in the past with this
aim.
OCR is used in many applications
li!e Resume processing, many
conversion applications, *ibrary
archives(digital library +covert large
boo! collections for on,line viewing
of content-
.ocument identification, ",boo!,
#nvoice and shipping receipt
processing, Phoneboo! processing,
%orensic analysis of handwriting.
Postal automation, Reading aid for
blind, *anguage processing, an!
Transaction documents processing
etc.
"$perimental .etails "$perimental .etails
/ardware Requirements0
) processor +Pentium 1 or higher-.
R)2 of 3 4 or higher.
/ard dis! +35 4 or higher-.
6oftware Requirements0
Operating system0 7indows 8P or 7indows 9.
2)T*)0 version :535 or higher.


Preprocessing of Handwritten Kannada Documents
asavantareddy;32633#615:<, Raghavendra 6wami .odamani;32633#6133<, asavantareddy;32633#615:<, Raghavendra 6wami .odamani;32633#6133<,
Prasanna 'umar;32633#6135<, )bhishe!;32635#6553< Prasanna 'umar;32633#6135<, )bhishe!;32635#6553<
Department of Information Science and Engineering
Place of work: MSRIT, Bangalore
Guide: Rajaram M Gowda
M.S.RAMAIAH INSTITUTE OF TECHNOLOGY M.S.RAMAIAH INSTITUTE OF TECHNOLOGY
Autonomous Institute Affiliated to VTU Autonomous Institute Affiliated to VTU! !
Open day - Project E!"#"t"on $%&' Open day - Project E!"#"t"on $%&'
MSRIT
MSRIT
M. S. .Ramaiah Institute of Technology M. S. .Ramaiah Institute of Technology M. S. .Ramaiah Institute of Technology M. S. .Ramaiah Institute of Technology
2ethodology ( Process 2ethodology ( Process
#mage enhancement is done using
4aussian blur filter, *aplace filter and
=62 +=nsharp 2as!ing- filter.
Then enhanced image is binarized
using Otsu algorithm in order to
differentiate bac!ground and
foreground of the ancient document.
6egmentation is carried out on present
handwritten 'annada documents using
connected component labeling and
bounding bo$ technique to segment the
individual characters

You might also like