Dynamic Bayesian networks for handwritten Arabic word recognition
| Autor: | Nabil Ghanmi, Amhad-Montaser Awal, Nihel Kooli |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Computer science
business.industry Speech recognition Feature extraction ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Computer Science::Computation and Language (Computational Linguistics and Natural Language and Speech Processing) Pattern recognition Image (mathematics) Set (abstract data type) ComputingMethodologies_PATTERNRECOGNITION Handwriting recognition Computer Science::Computer Vision and Pattern Recognition Word recognition Artificial intelligence Hidden Markov model business Word (computer architecture) Dynamic Bayesian network |
| Zdroj: | ASAR |
| DOI: | 10.1109/asar.2017.8067769 |
| Popis: | We present in this paper an approach for handwritten Arabic words recognition. A bi-dimensional model is created by coupling two Hidden Markov Models (HMM) using dynamic Bayesian network formalism. The image of the handwritten word is scanned in order to obtain horizontal and vertical information streams. Each information stream is associated to a HMM. Thus, the proposed coupling method allows taking into consideration both information sources to improve the word recognition. The proposed method is tested on a set of word from the IFN/ENIT database and the obtained result are very promising. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|