IJSRST1738150 | Received : 25 Nov 2017 | Accepted : 18 Dec 2017 | November-December-2017 [(3)8: 930-937]
© 2017 IJSRST | Volume 3 | Issue 8 | Print ISSN: 2395-6011 | Online ISSN: 2395-602X
Themed Section: Science and Technology
962
Analysis of Different Recognition Technique Classification and their Accuracy
Measure over Odia Script-A Review
Rajat Ketan Sahu
*
, Anita Minz, Sagarika Mishra, Chandrasekhar Panda
Department of Computer Science and Application, Sambalpur University, Jyoti Vihar, Burla, Odisha, India
rajat.keten.sahu@gmail.com, dr.chandrasekharpanda@gmail.com
ABSTRACT
This paper presents a literature review on character recognition techniques. (OCR) Optical Character Recognition
frameworks are able to identify varieties of text styles and printed characters. This means that by scanning a
document the device is able to essentially „read‟ the content. Character identification is a requesting specialization of
pattern recognition, machine learning, and advance image processing. Various work has been done as such far for
independent Indian language like Hindi, Urdu, Gujarati, Punjabi and onward. The latest research in this area has
been capable to grow some new methodologies to overcome the difficulty of Odia writing style. Be that as it may,
narrow exploration has been done over on odia content. Thus far various techniques are projected for segmentation,
feature extraction, classification, and recognition tasks of Odia characters. This paper tries to review, their relative
strength and weakness unit layout done by various research worker over on Odia content in last few years. The basic
objective of this paperwork is to nearby present an overview of different existing methods that have been developed
to recognize Odia character throughout last decade. The paper may be a concern the individuals who are interested
to work in the fields of acknowledgment of Odia script.
Keywords : Optical Character Recognition (OCR), Odia Script, Classification, Recognition.
I. INTRODUCTION
Optical character recognition (OCR) deal with the
technique of converting or changing over the
handwritten text or printed content into a machine
representation, that is understood by the machines for
easy editing and looking. The greatest trail inside the
field of image processing is to acknowledged the
documents both in machine printed and handwritten
structure [1]. Optical Character Recognition is a type of
document image analysis where a scanned digital image
that involves either machine printed or handwritten
script input into an OCR software engine converts it into
an editable machine readable digital text
format[2].Development of OCR for different Indian
languages is an active community of research work in
today‟s era. We are creating and aim to develop the
Hand written Character recognition framework for Odia
conversation, which is the official language of Odisha.A
part of research, is done on the printed Odia script,
however a little mass of work has been done on the
handwritten Odia content for recognition. Partition of
handwritten or printed text into lines, words, and
characters is one of the prime footsteps in the
handwritten text recognition outlook. The text of hand
written can be difficult to recognize due to variation in
font sizes, type of noises, color, contrast, alignment, and
background information. This kind of modification is
difficult to identify the word in the document while
writing [3]. Since handwritten text will vary on the
user‟s skill, orientation and manner of the writing
method. An in-depth study of Odia character recognition,
outlining existing ways and their drawbacks have been
carried out in this paper and future research direction in
this field has been suggested. The various steps of
character recognition such as image acquisition,
preprocessing, segmentation, feature extraction, and
post-processing.
There are collections of application domain where, OCR
can help. Major areas are represented below:
1. Protect past records in electronic format.
2. Save document pictures inside limited space.