Omnifont text recognition of printed cursive scripts via HMMs, compact lossless features, and soft data clustering.
Mohamed AttiaMohamed S. El-MahallawyMohsen A. RashwanWaleed NazihMohamed Al-BadrashinyPublished in: Pattern Anal. Appl. (2015)
Keyphrases
- data clustering
- text recognition
- optical character recognition
- unsupervised feature selection
- hidden markov models
- viterbi algorithm
- cluster analysis
- k means
- clustering algorithm
- unsupervised learning
- feature vectors
- image features
- feature set
- feature space
- character recognition
- feature extraction
- knn
- test images
- image sequences
- decision trees
- computer vision