Semisupervised Classification of PolSAR Image Incorporating Labels' Semantic Priors.
Biao HouJiaojiao GuanQian WuLicheng JiaoPublished in: IEEE Geosci. Remote. Sens. Lett. (2020)
Keyphrases
- image classification
- image features
- multiscale
- confidence scores
- input image
- single image
- image content
- class labels
- high resolution
- semantic labels
- image retrieval
- image analysis
- prior probabilities
- labeling process
- classification accuracy
- image data
- image collections
- bayesian framework
- machine learning
- prior information
- edge detection
- image representation
- segmented images
- prior model
- image segmentation
- training set
- similarity measure
- segmentation algorithm
- segmentation method
- text classification
- semantic segmentation
- weakly labeled
- auto annotation
- feature selection
- high level
- training data
- pairwise
- multi label classification
- feature extraction
- labeling problems
- support vector
- feature space
- prior knowledge
- training examples
- scene classification
- low level
- extracted features
- support vector machine
- semi supervised
- markov random field
- visual features
- multi label