275 Hits in 6.1 sec

Fast Multi-language LSTM-based Online Handwriting Recognition [article]

Victor Carbune and Pedro Gonnet and Thomas Deselaers and Henry A. Rowley and Alexander Daryin and Marcos Calvo and Li-Lun Wang and Daniel Keysers and Sandro Feuz and Philippe Gervais
2020 arXiv   pre-print
We describe an online handwriting system that is able to support 102 languages using a deep neural network architecture.  ...  This new system has completely replaced our previous Segment-and-Decode-based system and reduced the error rate by 20%-40% relative for most languages.  ...  We thank Google's OCR team for the numerous collaborations throughout the years that have made this work easier, as well as the speech recognition and machine translation teams at Google for tools and  ... 
arXiv:1902.10525v2 fatcat:xjp56djpzbfezf63lgqoxnzsvq

Online handwritten Gurmukhi word recognition using fine-tuned Deep Convolutional Neural Network on offline features

Sukhdeep Singh, Anuj Sharma, Vinod Kumar Chauhan
2021 Machine Learning with Applications  
Although, online handwriting recognition is a mature but exciting and fast developing field of pattern recognition, the same is not true for many of the Indic scripts.  ...  The previous Gurmukhi word recognition system followed the stroke based class labeling approach, whereas the present study has followed the word based class labeling approach.  ...  Multi-language online handwriting recognition based on beta-elliptic model and hybrid tdnn-svm classifier.  ... 
doi:10.1016/j.mlwa.2021.100037 fatcat:ohu2dktiifft3lyecf6g3yuauu

A hypothesize-and-verify framework for Text Recognition using Deep Recurrent Neural Networks [article]

Anupama Ray, Sai Rajeswar, Santanu Chaudhury
2015 arXiv   pre-print
Deep LSTM is an ideal candidate for text recognition.  ...  Thus the verification framework using language models eliminates wrong segmentation outputs and filters recognition errors.  ...  LSTM based approaches have Candidate word segments from each segmentation algorithm is outperformed HMM based ones for handwriting recognition passed through the Deep BLSTM recognizer  ... 
arXiv:1502.07540v1 fatcat:lmptvd5wxfcmhddkhh5bkn44ue

Handwriting Recognition Based on 3D Accelerometer Data by Deep Learning

Pedro Lopez-Rodriguez, Juan Gabriel Avina-Cervantes, Jose Luis Contreras-Hernandez, Rodrigo Correa, Jose Ruiz-Pinales
2022 Applied Sciences  
Online handwriting recognition has been the subject of research for many years. Despite that, a limited number of practical applications are currently available.  ...  This paper proposes a handwritten character recognition system based on 3D accelerometer signal processing using Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM).  ...  In order to attract the user's attention to the pen interface, fast and accurate handwriting recognition interfaces are highly required.  ... 
doi:10.3390/app12136707 fatcat:i6kxikpvxjbq7geaq5zneiazry

Handwritten Optical Character Recognition (OCR): A Comprehensive Systematic Literature Review (SLR) [article]

Jamshed Memon, Maira Sami, Rizwan Ahmed Khan
2020 arXiv   pre-print
Given the ubiquity of handwritten documents in human transactions, Optical Character Recognition (OCR) of documents have invaluable practical worth.  ...  Optical character recognition is a science that enables to translate various types of documents or images into analyzable, editable and searchable data.  ...  [9] demonstrated a template-based system for online character recognition, which is capable of representing different handwriting styles of a particular character.  ... 
arXiv:2001.00139v1 fatcat:p3rdutz35besxfxf7suozt7r2u

Derin Öğrenme Araştırma Alanlarının Literatür Taraması

M. Mutlu Yapıcı, Adem Tekerek, Nurettin Topaloğlu
2019 Gazi Mühendislik Bilimleri Dergisi  
In the last decade, the-state-of-the-art studies on many research areas such as computer vision, object recognition, speech recognition, and natural language processing were especially led to the awakening  ...  Çalışmada Özerk Araçlar (Autonomous Vehicles), Doğal Dil İşleme (Natural Language Processing), El Yazısı Karakter Tanıma (Handwritten Character Recognition), İmza Doğrulama (Signature Verification), Ses  ...  Handwriting recognition is composed of two types such online (stroke trajectory-based) and offline (image.-based).  ... 
doi:10.30855/gmbd.2019.03.01 fatcat:2sv7dg7elrfqppcjx5otzmb7pi

Real-Time Pashto Handwritten Character Recognition Using Salient Geometric and Spectral Features

Muhammad Shabir, Naveed Islam, Zahoor Jan, Inayat Khan, Taj Rahman, Asim Zeb, Shafiq Ahmad, Abdelaty Edrees Sayed, Mali Abdollahian
2021 IEEE Access  
For real-time recognition of characters, the probability-based multi-class Naïve Bayesian classifier is exploited, which determines the probabilities of geometric invariant features to predict the character  ...  The performance of the proposed approach has been validated through extensive experiments and based on the recognition matrices, the proposed technique achieves an accuracy of 97.5% for online Pashto handwritten  ...  The existing online or real-time handwriting recognition systems based on the Naive Bayes technique are computationally less expensive as compared with ANN and SVM-based techniques.  ... 
doi:10.1109/access.2021.3123726 fatcat:ghs75tntczezxf275ehoyzrjpm

Exploring Deep Learning Approaches to Recognize Handwritten Arabic Texts (May 2020)

Mohamed Eltay, Abdelmalek Zidouri, Irfan Ahmad
2020 IEEE Access  
In this paper, we review and investigate different deep learning architectures and modeling choices for Arabic handwriting recognition.  ...  This weight is calculated based on the average probability of each class in a word.  ...  HANDWRITING RECOGNITION SYSTEMS In general, character recognition systems can be divided into two types, as shown in Fig.7 : online and offline.  ... 
doi:10.1109/access.2020.2994248 fatcat:6cdj2t5jmnccnjjy6grq734ddm

Deep Learning: Our Miraculous Year 1990-1991 [article]

Juergen Schmidhuber
2021 arXiv   pre-print
Back then, few people were interested, but a quarter century later, neural networks based on these ideas were on over 3 billion devices such as smartphones, and used many billions of times per day, consuming  ...  [JOU17] Google's new on-device speech recognition of 2019 (now on your phone, not on the server) is still based on LSTM.  ...  [NAS] Since 2006, we have worked with the software industry (e.g., LifeWare) to greatly improve handwriting recognition.  ... 
arXiv:2005.05744v2 fatcat:5udqp5zwzfd7veedfpqrvsathe

Recurrence-free unconstrained handwritten text recognition using gated fully convolutional network [article]

Denis Coquenet, Clément Chatelain, Thierry Paquet
2020 arXiv   pre-print
Unconstrained handwritten text recognition is a major step in most document analysis tasks.  ...  In this paper we present a Gated Fully Convolutional Network architecture that is a recurrence-free alternative to the well-known CNN+LSTM architectures.  ...  In [4] , a Multi-Dimensional LSTM (MDLSTM) is used to manage dependencies over both horizontal and vertical axis.  ... 
arXiv:2012.04961v1 fatcat:z7mx2kkavjhi5j23juxk4uqrnu

Transcript Anatomization with Multi-Linguistic and Speech Synthesis Features

Rohan Modi
2021 International Journal for Research in Applied Science and Engineering Technology  
Handwriting Recognition is the process of transforming a handwritten text in a specific language into its digitally expressible script represented by a set of icons known as letters or characters.  ...  While there are many systems which convert normal language text in to speech, the aim of this paper is to study Optical Character Recognition with speech synthesis technology and to develop a cost effective  ...  Offline handwriting recognition is relatively difficult, as different people have different handwriting styles.  ... 
doi:10.22214/ijraset.2021.35371 fatcat:tbtekzwpergqzonrl4n2ag6u3q

Recent Advances in Recurrent Neural Networks [article]

Hojjat Salehinejad, Sharan Sankar, Joseph Barfett, Errol Colak, Shahrokh Valaee
2018 arXiv   pre-print
In [122] , a BLSTM model is introduced for online handwriting recognition.  ...  Jaeger Leaky integration neurons 2007 Graves MDRNN: Multi-dimensional RNNs 2009 Graves LSTM for hand-writing recognition 2010 Mikolov RNN based language model 2010 Neir Rectified linear  ... 
arXiv:1801.01078v3 fatcat:ioxziqbkmzdrfoh2kukul6xlku

Robust Handwriting Recognition with Limited and Noisy Data [article]

Hai Pham, Amrith Setlur, Saket Dingliwal, Tzu-Hsiang Lin, Barnabas Poczos, Kang Huang, Zhuo Li, Jae Lim, Collin McCormack, Tam Vu
2020 arXiv   pre-print
Despite the advent of deep learning in computer vision, the general handwriting recognition problem is far from solved.  ...  Most existing approaches focus on handwriting datasets that have clearly written text and carefully segmented labels.  ...  Fast r-cnn.  ... 
arXiv:2008.08148v1 fatcat:4uu2d5ncsjduvjwo2ejfuikb2y

Amharic OCR: An End-to-End Learning

Birhanu Belay, Tewodros Habtegebrial, Million Meshesha, Marcus Liwicki, Gebeyehu Belay, Didier Stricker
2020 Applied Sciences  
In this paper, we introduce an end-to-end Amharic text-line image recognition approach based on recurrent neural networks.  ...  The experimental results, on a printed and synthetic benchmark Amharic Optical Character Recognition (OCR) database called ADOCR, demonstrated that the proposed model outperforms state-of-the-art methods  ...  unconstrained online handwriting recognition [34] , MDLSTM for handwriting recognition [35] , an online handwritten mathematical expression recognition using a Gated Recurrent Unit (GRU)-based attention  ... 
doi:10.3390/app10031117 fatcat:4gtjsm5gcjggtlvfsymqurapgm

Attention-based Fully Gated CNN-BGRU for Russian Handwritten Text [article]

Abdelrahman Abdallah, Mohamed Hamada, Daniyar Nurseitov
2020 arXiv   pre-print
This research approaches the task of handwritten text with attention encoder-decoder networks that are trained on Kazakh and Russian language.  ...  We developed a novel deep neural network model based on Fully Gated CNN, supported by Multiple bidirectional GRU and Attention mechanisms to manipulate sophisticated features that achieve 0.045 Character  ...  The following section investigates the related work on Offline Handwriting Text Recognition. Section three demonstrates the attention-based fully gated convolutional recurrent neural network.  ... 
arXiv:2008.05373v5 fatcat:msf4nhqcwnhbfokz2ui3gp4i3i
« Previous Showing results 1 — 15 out of 275 results