SAR Image Classification Using Fully Connected Conditional Random Fields Combined with Deep Learning and Superpixel Boundary Constraint.
Zhensheng SunMiao LiuPeng LiuJuan LiTao YuXingfa GuJian YangXiaofei MiWeijia CaoZhouwei ZhangPublished in: Remote. Sens. (2021)
Keyphrases
- conditional random fields
- fully connected
- deep learning
- restricted boltzmann machine
- sar images
- superpixels
- unsupervised learning
- hidden markov models
- graphical models
- probabilistic model
- maximum entropy
- long range
- random fields
- machine learning
- higher order
- information extraction
- weakly supervised
- markov random field
- pattern recognition
- pairwise
- named entity recognition
- generative model
- supervised learning
- decision trees
- feature space
- support vector
- training samples
- image classification
- data mining
- approximate inference
- image features
- image processing
- learning algorithm
- feature vectors