IJSRSET162237 | Received: 02 March 2016 | Accepted:20 March 2016 | March-April2016 [(2)2: 157-160]
© 2016 IJSRSET | Volume 2 | Issue 2 | Print ISSN : 2395-1990 | Online ISSN : 2394-4099
Themed Section: Engineering and Technology
157
A Survey on Gujarati Handwritten OCR using Morphological Analysis
Vaidehi Patel, Prof. Abhinay Pandya
Department of Computer Engineering, LDRP Institute of Technologya and Research, Gandhinagar,Gujarat, India
ABSTRACT
Handwriting recognition has been one of the active and challenging research areas in the field of imageprocessing
and pattern recognition. Handwriting recognition can be differentiated into two categories i.e.online handwriting
recognition and Offline handwriting recognition. I mainly focus on offline handwritingrecognition. It is also
observed that majority of work for this language is for printed form rather thanhandwritten form. One can obtain
very less OCR related research work for Gujarati script, especially forhandwritten form. I use Freeman chain code
for feature extraction and then used Hidden Markov method.There is no benchmark dataset is available so it is
generated by collecting sample from more than 100writers of different age group and gender.
Keywords: Handwritten Recognition, Freeman Chain Code, Pre-processing, Classification
I. INTRODUCTION
The form of the text is varied from the scanneddocument
to the typed text in various fonts. For
humans,recognition of text is a trivial task, but to make
amachine that recognizes characters is extremelydifficult.
The current study is being focused onexploration of
possible methods to develop an OCRsystem for Gujarati
language when noise is present inthe signal. In order to
understand the core challenges, athorough analysis of
Gujarati Writing System has beendone. To know and
understand the research in this field,existing OCR
systems are studied. The Prominence wason finding
workable segmentation technique anddiacritic handling
for Gujarati words, and building arecognition module for
these ligatures.
To develop an OCR system Gujarati text, a
completeprocedure is proposed as well as a testing
application isalso made. Test results from this are
reported andcompared with the previous work done in
this area.
II. METHODS AND MATERIAL
A. Features of Gujarati Language
Gujarati is regional language of state Gujarat in India.
Gujarati is name of script which is written and spokenby
people in Gujarat. Gujarati script has 34+2 constantsand
11 vowels. The structures of some Gujaratiharacters are
very similar to Devnagri script. Devngari script has
shirolekha on top of the characters, but Gujarati script has
not. Gujarati script does not have the distinction of Lower
and Upper Cases like English script. Gujarati script has
combination of constants and vowels. Every vowel has a
unique symbol, called vowels modifiers. Table 1 shows
Gujarati vowels, consonants, matras, other symbols and
some of the conjuncts [5].
Figure 1 : Gujarati Script[5]