Recognizing handwritten Arabic words using grapheme segmentation and recurrent neural networks.
Gheith A. AbandahFuad T. JamourEsam A. Al QarallehPublished in: Int. J. Document Anal. Recognit. (2014)
Keyphrases
- recurrent neural networks
- numeral strings
- arabic documents
- cursive script
- word segmentation
- word recognition
- handwritten words
- handwritten documents
- handwriting recognition
- neural network
- newspaper articles
- feed forward
- word spotting
- reservoir computing
- echo state networks
- text lines
- feedforward neural networks
- segmentation algorithm
- neural model
- arabic language
- image segmentation
- character segmentation
- recurrent networks
- level set
- medical images
- short text
- unknown words
- cascade correlation
- character recognition
- nonlinear dynamic systems
- arabic handwriting recognition
- document analysis
- automatic recognition
- segmentation method
- n gram
- optical character recognition
- sign recognition
- artificial neural networks
- artificial intelligence
- writer identification
- historical documents
- morphological analysis
- printed text
- arabic text