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
TL;DR: Javanese calligraphy (hanacaraka) is a derivative of the Devanagari letters, also known as hanacaraaka as mentioned in this paper , which is used in various manuscripts in the form of chronicles, poetry and other literary writings.
Abstract: Javanese calligraphy (hanacaraka) is a derivative of the Devanagari letters, also known as hanacaraka. Javanese script is estimated to have started
to be used in the Islamic Mataram era in 1608. This letter is used in various manuscripts in the form of chronicles, poetry and other literary writings.
There has been no writing in the form of a review related to the potential that can be used for therapy. The users of this letter are Javanese, the
number of Indonesian people and it is ofcially taught at the elementary & junior high school level in the provinces of Central Java and East Java.
We will describe the forms of Javanese letters and their potential use from the Neuroaesthetic side.
01 Jan 2023
TL;DR: In this article , a convolutional neural network (CNN) architecture was proposed to identify six word-level handwritten scripts involving Arabic, Latin, Chinese, Bangla, Devanagari and Telugu.
Abstract: In this work, we propose a convolutional neural network (CNN) architecture to identify six word-level handwritten scripts involving Arabic, Latin, Chinese, Bangla, Devanagari and Telugu. A large dataset of 14k word images per script was constructed based on several public handwritten datasets. Then, three architectures are proposed and compared based on standard metrics performance and time execution. Experiments conducted on both test and validation classification show high performances that outperform the state-of-art techniques. Indeed, the best result was provided by CNN model with three-convolutional-polling pairs layers that achieved an average script identification accuracy of 97.67% and ran in a sufficiently fast time of 2 ms per frame during the test phase.
TL;DR: In this paper , a sample of images which are centralized and grayscale are considered and analyzed using the K-nearest neighbor classification, extremely randomized decision forest classification and random forest classification are considered.
Abstract: India is a country of various languages right from Kashmir to Kanyakumari. The national language of India is Hindi which is also the third most popular language in the world. The script in which the Hindi language is written is known as Devanagari script which in fact is used to write many other languages such as Sanskrit, Marathi, Nepali, and Konkani languages. Neural networks are recently being used in several different ways of pattern identification. It is common knowledge that every person’s handwriting is dissimilar. Therefore, it is challenging to recognize those handwritten monograms. The sector of pattern recognition that has become a hot topic for research purposes is handwritten character recognition. This is where neural networks play an important role. The competence of a computer to take in and decipher comprehensible transcribed input whose origin is paper documents, touch screens, photographs, and alternative gadgets are termed as handwriting recognition. Handwritten recognition of words is a model which is used to convert the written text into words that are crucial in the human computer interface. The handwriting recognition area is an extensively experimented branch till date and the Devanagari script recognition is progressing area of research. The above application is used in mail sorting, office automation, cheque verification, and human computer communication, i.e. the growing age of artificial intelligence. A sample of the dataset of images which are centralized and grayscale are considered and analysed using the K-nearest neighbour classification, extremely randomized decision forest classification and random forest classification are considered.
Posted Content•
TL;DR: In this paper, the authors focus on analyzing hate speech in Hindi-English code-switched language and propose a Map Only Hindi (MoH) method to contain the structure of data and use it with existing algorithms.
Abstract: Social media has become a bedrock for people to voice their opinions worldwide. Due to the greater sense of freedom with the anonymity feature, it is possible to disregard social etiquette online and attack others without facing severe consequences, inevitably propagating hate speech. The current measures to sift the online content and offset the hatred spread do not go far enough. One factor contributing to this is the prevalence of regional languages in social media and the paucity of language flexible hate speech detectors. The proposed work focuses on analyzing hate speech in Hindi-English code-switched language. Our method explores transformation techniques to capture precise text representation. To contain the structure of data and yet use it with existing algorithms, we developed MoH or Map Only Hindi, which means "Love" in Hindi. MoH pipeline consists of language identification, Roman to Devanagari Hindi transliteration using a knowledge base of Roman Hindi words. Finally, it employs the fine-tuned Multilingual Bert and MuRIL language models. We conducted several quantitative experiment studies on three datasets and evaluated performance using Precision, Recall, and F1 metrics. The first experiment studies MoH mapped text's performance with classical machine learning models and shows an average increase of 13% in F1 scores. The second compares the proposed work's scores with those of the baseline models and offers a rise in performance by 6%. Finally, the third reaches the proposed MoH technique with various data simulations using the existing transliteration library. Here, MoH outperforms the rest by 15%. Our results demonstrate a significant improvement in the state-of-the-art scores on all three datasets.
01 Jul 2020
TL;DR: Thai handwritten text recognition performance is considerably improved by multi-task learning, and the model consists of three layers employing Convolutional Neural Network, Recurrent Neural Network and Connectionist Temporal Classification, respectively.
Abstract: Written languages have some commonalities and differences. Knowing one language can help learning other languages better. While labelled data is limited in offline Thai handwritten text recognition problem domain, multi-task learning is chosen to address the problem in our study. The multi-task model is trained from Thai, Latin and Devanagari writing scripts. The model consists of three layers employing Convolutional Neural Network, Recurrent Neural Network and Connectionist Temporal Classification, respectively. Recognition accuracies are compared against three corresponding single-task models. Thai handwritten text recognition performance is considerably improved by multi-task learning. Learning multiple languages helps the model generalize better when trained by large enough datasets.