Segmentation of Printed Devnagari Documents

Dongre, Vikas J.; Mankar, Vijay H.

doi:10.1007/978-3-642-22555-0_23

Cited by 5 publications

(2 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is limited research addressing this problem: methods based on HMM which generally aimed at extracting feature vectors based on a sliding window (Elzobi, Al-Hamadi, Dings, Elmezain, & Saeed, 2013;Slimane, Ingold, Kanoun, Alimi, & Hennebert, 2010;Tz & Fink, 2009), SVM (Khalifa & BingRu, 2011), or neural network (Al-Jawfi, 2009). Other methods are based on transforms such as the Hough or wavelet transforms (Touj, Amara, & Amiri, 2007), strokes, segments and tokens (Cheung, Bennamoun, & Bergmann, 2001;El-Dabi, Ramsis, & Kamel, 1990), contour tracing (Bushofa & Spann, 1997;Romeo-Pakker, Miled, & Lecourtier, 1995), histogram, and baseline (Amara & Ellouze, 1995;Dongre, Vikas, & Mankar, 2011). A survey of Arabic character segmentation can be found in (Alginahi, 2013).…”

Section: Ahl Methodsmentioning

confidence: 99%

Trained convolutional neural network based on selected beta filters for Arabic letter recognition

ElAdel

Zaied

Amar

2018

WIREs Data Min & Knowl

View full text Add to dashboard Cite

This paper presents a fast deep learning approach to segment and recognize off‐line Arabic printed and handwritten letters from words. We proposed a simple and powerful algorithm for Arabic letter segmentation based on vertical profile and baseline analysis. Then, we proposed a new method for feature extraction using fast wavelet transform. These extracted features are exploited as connection weights to build a convolutional neural network for each letter shape. Finally, all estimated model shapes are boosted to increase the robustness and performance of the proposed system. The proposed approach was tested on APTI and IESK‐arDB databases to evaluate performance for printed letters and handwritten letters, respectively. The obtained results show the robustness of our approach as well as the speed of the proposed recognition algorithm for both databases. This article is categorized under: Technologies > Machine Learning Technologies > Computational Intelligence Technologies > Classification

show abstract

Section: Ahl Methodsmentioning

confidence: 99%

Trained convolutional neural network based on selected beta filters for Arabic letter recognition

ElAdel

Zaied

Amar

2018

WIREs Data Min & Knowl

View full text Add to dashboard Cite

show abstract

“…The method of printed Devanagari document segmentation technique using profiling values was proposed (Dongre & Mankar, 2011), which achieved 100% line and word segmentation accuracy.RamanandSahu(2015)providedDevanagarisegmentationbyremovingshirorekhaand extracting vowels, consonants, lower and upper modifiers from words and achieved 87.73% and 86.4%accuraciesforcharactersandwords,respectively.Nextcharactersegmentationmethodused shape-orientedfeaturevectorsforprintedDevanagari(Jindal&Kumar2017)andachieved98.56% accuracy in the time efficient manner. In this direction, another method segmented the printed scene-basedDevanagariwords,intocharactersandobtained55.77%averageperformanceand4.76 sec.executiontime (Murthy,Roy,Narang,Hanmandlu&Gupta,2013).Italsodiscussedthepoor resultsobtainedinblurredimages,closelyhandwrittenimagesandover-segmentation.Garainand Chaudhuri(2002)developedapredictivealgorithmtosegmentthetouchingcharactersinprinted Devanagari and Bangla using cut columns selection with fuzzy multifactorial analysis.…”

Section: Thematic Backgroundmentioning

confidence: 99%

A Hybrid Hindi Printed Document Classification System Using SVM and Fuzzy

Puri

Singh

2019

Journal of Information Technology Research

View full text Add to dashboard Cite

This article introduces a new advanced tri-layered segmentation and bi-leveled-classifier-based Hindi printed document classification system, which categorizes imaged documents into pre-defined mutually exclusive categories by using SVM and Fuzzy matching at character and document classifications, respectively. During training, the improved and noise-free image is segmented into lines and words by profiling. Then it obtains Shirorekha Less (SL) isolated characters along with upper, left and right modifier components from the SL words. These components use their locations and inter character-modifier component distance to get associate with their corresponding characters only. Further, confidence values of all characters are calculated with SVM training and all characters are mapped into Romanized labels to generate the words. Finally, documents are classified by Fuzzy based matching of Romanized detected words and predefined classes. The average execution times of SL characters are 0.22675 sec. and 0.20375 sec. and classification accuracy are 74.61% and 80.73% for training and testing, respectively.

show abstract

A novel local skew correction and segmentation approach for printed multilingual Indian documents

Soora

Deshpande

2018

Alexandria Engineering Journal

View full text Add to dashboard Cite

Segmentation of Printed Devnagari Documents

Cited by 5 publications

References 9 publications

Trained convolutional neural network based on selected beta filters for Arabic letter recognition

Trained convolutional neural network based on selected beta filters for Arabic letter recognition

A Hybrid Hindi Printed Document Classification System Using SVM and Fuzzy

A novel local skew correction and segmentation approach for printed multilingual Indian documents

Contact Info

Product

Resources

About