The RWTH Large Vocabulary Arabic Handwriting Recognition System
| Autor: | Michal Kozielski, Mahdi Hamdani, Hermann Ney, Patrick Doetsch, Amr El-Desoky Mousa |
|---|---|
| Rok vydání: | 2014 |
| Předmět: |
Vocabulary
business.industry Computer science Intelligent character recognition Speech recognition media_common.quotation_subject Feature extraction computer.software_genre ComputingMethodologies_PATTERNRECOGNITION Recurrent neural network Discriminative model Handwriting recognition ComputingMethodologies_DOCUMENTANDTEXTPROCESSING Language model Artificial intelligence business Hidden Markov model computer Natural language processing media_common |
| Zdroj: | Document Analysis Systems |
| DOI: | 10.1109/das.2014.61 |
| Popis: | This paper describes the RWTH system for large vocabulary Arabic handwriting recognition. The recognizer is based on Hidden Markov Models (HMMs) with state of the art methods for visual/language modeling and decoding. The feature extraction is based on Recurrent Neural Networks (RNNs) which estimate the posterior distribution over the character labels for each observation. Discriminative training using the Minimum Phone Error (MPE) criterion is used to train the HMMs. The recognition is done with the help of n-gram Language Models (LMs) trained using in-domain text data. Unsupervised writer adaptation is also performed using the Constrained Maximum Likelihood Linear Regression (CMLLR) feature adaptation. The RWTH Arabic handwriting recognition system gave competitive results in previous handwriting recognition competitions. The used techniques allows to improve the performance of the system participating in the OpenHaRT 2013 evaluation. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|