Algorithm for segmenting script-dependant portion in a bilingual Optical Character Recognition system

Bebartta, Himadri Nandini Das; Mohanty, Sanghamitra

doi:10.1134/s1054661817030142

Cited by 7 publications

(3 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Now the segments of the test page which are Telugu characters are given to the trained model for prediction. Later SoftMax is applied to classify images, which uses the expression to generate discrete probabilities of all classes between 0 and 1 given by (8),…”

Section: Character Segmentation and Feature Extractionmentioning

confidence: 99%

See 1 more Smart Citation

Computationally efficient handwritten Telugu text recognition

Revathi,

Prasad,

Gattim

2024

IJEECS

View full text Add to dashboard Cite

<p>Optical character recognition (OCR) for regional languages is difficult due to their complex orthographic structure, lack of dataset resources, a greater number of characters and similarity in structure between characters. Telugu is popular language in states of Andhra and Telangana. Telugu exhibits distinct separation between characters within a word, making a character-level dataset sufficient. With a smaller dataset, we can effectively recognize more words. However, challenges arise during the training of compound characters, which are combinations of vowels and consonants. These are considered as two or more characters based on associated vattus and dheerghams with the base character. To address this challenge, each compound character is encoded into a numerical value and used as input during training, with subsequent retrieval during recognition. The segmentation issue arises from overlapping characters caused by varying handwritten styles. For handling segmentation issues at the character level arising from handwritten styles, we have proposed an algorithm based on the language's features. To enhance word-level accuracy a dictionary-based model was devised. A neural network utilizing the inception module is employed for feature extraction at various scales, achieving word-level accuracy rates of 78% with fewer trainable parameters.</p>

show abstract

Section: Character Segmentation and Feature Extractionmentioning

confidence: 99%

“…Third phase is character segmentation, a crucial step in OCR [7], [8]. This phase depends on language.…”

Section: Introductionmentioning

confidence: 99%

Computationally efficient handwritten Telugu text recognition

Revathi,

Prasad,

Gattim

2024

IJEECS

View full text Add to dashboard Cite

show abstract

“…( 2011)presentedarobustandfastword-wiseidentificationmethodbyisolatingpages, blocks,andparagraphsfromreal-worldbi-lingualdatasets.Theirmethodperformedpreprocessing first,thenfoundtexturefeaturesatpageandblocklevelsandstructuralfeaturesatwordlevel,then usedprofilebasedsegmentationforblocksandwords,andfinally,performedSVMclassification atpageandblocklevelsandRejectionbasedclassificationusingAdaBoostatthewordlevel.The scriptidentificationmethodproposedbyRani,Dhir,andLehal(2012)recognizedbi-scriptwords through preprocessing, feature extraction of structural, Gabor and Discrete Cosine Transforms (DCT), and finally, classification by SVM, K Nearest Neighbor (KNN) and Probabilistic Neural Network(PNN).Nextword-levelidentificationtechniqueclassifiedbi-scriptdocumentsbyusing worddirectionalenergydistributionfeaturesofGaborfiltersalongwithsuitablefrequenciesand orientations (Chaudhari & Gulati, 2016). Bebartta and Mohanty (2017) 2018)reviewedwordrecognitiontechniquesforIndicandnon-Indicscripts bydiscussingtheirexperimentalresults,databases,recognitionaccuracies,potentialbenefits,and future recommendations for Indic scripts, such as Bengali, Devanagari, Gujarati, Gurumukhi, Kannada,Maithili,Malayalam,Oriya,Tamil,andTelugu,andnon-Indicscripts,suchasArabic, Chinese,Dutch,Japanese,Latin/Roman,Mongolian,Persian,Thai,andUyghur.Theydiscussedword recognitiontypes,approaches,needs,advantages,disadvantages,issues,andchallengesalongwith surveyprotocols,theirdevelopment,conduct,resultanalysis,reporting,andfindings.Theyfound theneedofadvancedprinted/handwrittendocumentrecognitiontechniquesandwordsegmentation algorithmstoachievehighwordrecognitionaccuracy.GhoshandValveny(2018)proposedafast, segmentation-free,wordspottingmethod,andfollowedthestepsofatomicboundingboxgeneration tocreatetextboxandfilterproposals,PyramidalHistogramofCharacters(PHOC)featureencoding toevaluatetheseproposals,performedindexingbyusingPyramidalHistogramofCharacterN-grams (PHON),andfinally,attributemodellearningbyLSVM.Theirmethodshowedtheperformance forquery-by-stringandquery-by-examplewithstandardsingleandmulti-writerdatasets.Table2 providesthecomparisonsamongwordrecognitionandspottingtechniquesfortheyearrange2014 to2018,whichdiscriminatesthesetechniqueswiththeindicatorsofprintedandhandwritten,script andlanguage,datasetandsize,classifierused,theaccuracyachieved,andconstraints,errors,and futuredirections.…”

Section: Script Identification and Recognitionmentioning

confidence: 99%

A Fuzzy Matching based Image Classification System for Printed and Handwritten Text Documents

Puri

Singh

2020

Journal of Information Technology Research

View full text Add to dashboard Cite

This article proposes a bi-leveled image classification system to classify printed and handwritten English documents into mutually exclusive predefined categories. The proposed system follows the steps of preprocessing, segmentation, feature extraction, and SVM based character classification at level 1, and word association and fuzzy matching based document classification at level 2. The system architecture and its modular structure discuss various task stages and their functionalities. Further, a case study on document classification is discussed to show the internal score computations of words and keywords with fuzzy matching. The experiments on proposed system illustrate that the system achieves promising results in the time-efficient manner and achieves better accuracy with less computation time for printed documents than handwritten ones. Finally, the performance of the proposed system is compared with the existing systems and it is observed that proposed system performs better than many other systems.

show abstract

Alignment and disruption robust binary mapper for optical Braille recognition

Juneja

Rana

2017

Int. j. inf. tecnol.

View full text Add to dashboard Cite

Algorithm for segmenting script-dependant portion in a bilingual Optical Character Recognition system

Cited by 7 publications

References 3 publications

Computationally efficient handwritten Telugu text recognition

Computationally efficient handwritten Telugu text recognition

A Fuzzy Matching based Image Classification System for Printed and Handwritten Text Documents

Alignment and disruption robust binary mapper for optical Braille recognition

Contact Info

Product

Resources

About