Dyadic Multi-resolution Analysis-Based Deep Learning for Arabic Handwritten Character Classification.
Asma ElAdelRidha EjbaliMourad ZaiedChokri Ben AmarPublished in: ICTAI (2015)
Keyphrases
- deep learning
- multi resolution analysis
- unsupervised learning
- handwritten characters
- machine learning
- pattern recognition
- restricted boltzmann machine
- classification accuracy
- supervised learning
- text classification
- learning algorithm
- feature space
- feature vectors
- handwritten character recognition
- data fusion
- unsupervised feature learning
- weakly supervised
- model selection
- markov random field
- semi supervised
- multiresolution
- decision making
- feature selection