Topic
Devanagari
About: Devanagari is a research topic. Over the lifetime, 655 publications have been published within this topic receiving 7428 citations. The topic is also known as: Deva nagari & Hindi Script.
Papers published on a yearly basis
Papers
More filters
03 Sep 2021
TL;DR: In this paper, a convolution neural network (CNN) was used for detecting and characterizing information of various dialects in light of its element extraction and preparing at various pooling and convolution stages.
Abstract: Finding Trust on Writer Identification assumes a significant part in misrepresentation discovery while thinking about the instance of unapproved access in banking and other privacy checks. Indic Forensic archive investigation is additionally done by trusted author identification (TWI). We have tested penmanship acknowledgment utilizing a convolution neural organization utilizing Tensor-Flow and contrast with discovering Trust. We have prepared and tried neural organization for transcribed archives of Different understudies. After effectively perceiving and characterizing the transcribed records we at that point prepared and tried the information of certain understudies for improving the exactness of the organization. It's extremely mind-boggling to effectively remove the words and characters from a transcribed report to getting the Trust written in Hindi (Devanagari). At that point, we utilized our organization ofother Indic Language like Kannada. The convolution neural organization is productive for perceiving or characterizing information of various dialects in light of its element extraction and preparing at various pooling and convolution stages. So It is powerful and trusted because it's chipping away at an alternate data set like Devanagari, Kannada, and Arabic data set with their confided in an incentive at various stages. Dealing with the security framework is certainly not an easy breezy. Contingent on the correspondence organization, the degree of security may differ. Those requiring a huge information sharing organization for the most part utilize a high-upkeep security framework alongside a serious programming/equipment framework that works in relationship with the security framework to successfully distinguish and keep assaults from the few malevolent offices. So wecheckthetrustindifferentwriteridentificationtechniques.
Patent•
31 May 2017
TL;DR: In this article, an image recognition method of converting a Sanskrit Devanagari printed character to Latin was proposed, which is easy to operate and high in efficiency, and is applicable to actual literature research and application.
Abstract: The invention discloses an image recognition method of converting a Sanskrit Devanagari printed character to Latin. The method comprises steps: (1) a character image containing a Sanskrit Devanagari printed character is scanned, segmentation of the Sanskrit Devanagari printed character is carried out based on the vertical maximum blank space among character blocks, and a plurality of Sanskrit Devanagari printed character blocks are obtained; (2) the obtained Sanskrit Devanagari printed character blocks are recognized to obtain feature vectors corresponding to the Sanskrit Devanagari printed character blocks; and (3) the obtained feature vectors and feature vectors of standard Latin characters are compared, and according to a comparison result, the recognized Sanskrit Devanagari printed character blocks are converted to a Latin character. Thus, direct conversion from the Sanskrit Devanagari noiseless printed character image to the corresponding Latin character can be realized, the accuracy is high and basically, 100% accuracy can be achieved. The used image recognition algorithm is convenient and easy to operate and high in efficiency. The technical scheme provided by the invention is easy to realize and is applicable to actual literature research and application.
10 Apr 2023
TL;DR: In this article , the authors proposed a bidirectional Encoder Representations from Transformers (BERT) based contextual embedding technique with a concatenation of emoji2vec Embedings to classify social media posts in Hindi Devanagari script as hostile or non-hostile.
Abstract: Detection of hostile content from social media posts (FacebookTM, TwitterTM etc.) is a demanding task in the field of Natural Language Processing (NLP). Daily growing nature of hostile content in different electronic media opened up new challenges in language understanding. It becomes more difficult in regional languages. AI-based solution is required to identify hostile content on a large scale. Though a satisfactory amount of researches has been carried out in the English language, finding hostile content in regional languages is still under progress due to unavailability of suitable datasets and tools. In terms of the number of speakers, Hindi ranks third in the world and first in the Indian Subcontinent. The objective of the article is to design hostile content detection system in Hindi language using coarse-grained (binary) classification and fine-grained (multi-class, multi-label) classification. We noted that different baseline learning method with different pre-trained language models perform differently. Using the Constraint 2021 Hindi Dataset, this research proposes a Bidirectional Encoder Representations from Transformers (BERT) based contextual embedding technique with a concatenation of emoji2vec Embedings to classify social media posts in Hindi Devanagari script as hostile or non-hostile. Additionally, for the fine-grained tasks where hostile posts are sub-categorized as defamation, fake, hate, and offensive, we develop an Ensemble Classifier varying different learning methods and embedding models. With an F1-Score of 0.9721, it is found that our proposed Indic-BERT+emoji model outperforms the baseline model and other existing models for the coarse-grained task. We have also observed that our proposed Ensemble method is giving good results than the existing models and the baseline model for the fine-grained tasks with F1-Score of 0.43, 0.82, 0.58 and 0.62 for defamation, fake, hate, and offensive classes respectively. The code and the data are available in https://github.com/skarifahmed/hostile.
TL;DR: Comparison study of various approaches for Handwritten Devanagari Character recognition using neural network using artificial neural network has been presented.
Abstract: Handwritten character can be converted in to the digital information using handwritten character Recognition, which is the ability of a computer to receive and interpret handwritten input from documents. Handwritten Devanagari Characters are more complex for recognition due to presence of header line, conjunct characters and similarity in shapes of multiple characters. For Handwritten Devanagari Character recognition using neural network various approaches has been proposed. In general the process involves phases as: Scanning, Preprocessing, Feature Extraction and Recognition. Preprocessing includes noise reduction, binarization, normalization and thinning. Feature extraction includes extracting some useful information out of the thinned image in the form of a feature vector. Artificial neural network is used for classification. In this survey, comparative study of various approaches has been presented.