Joint training of conditional random fields and neural networks for stroke classification in online handwritten documents.
Jun-Yu YeYan-Ming ZhangCheng-Lin LiuPublished in: ICPR (2016)
Keyphrases
- conditional random fields
- training process
- neural network
- structured prediction
- pattern recognition
- error correcting output coding
- sequence labeling
- handwritten documents
- graphical models
- higher order
- multiclass classification
- hidden markov models
- support vector
- training set
- probabilistic model
- named entity recognition
- information extraction
- markov random field
- feature vectors
- crf model
- machine learning
- training data
- feature extraction
- image processing
- computer vision
- pairwise
- supervised learning
- image classification
- generative model
- feature space
- superpixels
- web page prediction