Unsupervised Learning of Low Dimensional Satellite Image Representations via Variational Autoencoders.
Silvia ValeroFerran AgullóJordi IngladaPublished in: IGARSS (2021)
Keyphrases
- image representation
- unsupervised learning
- low dimensional
- dimensionality reduction
- feature space
- object recognition
- deep belief networks
- high dimensional
- multiscale
- high dimensional data
- image classification
- denoising
- bag of words
- principal component analysis
- supervised learning
- image content
- data points
- image retrieval
- feature representations
- sparse coding
- input space
- image features
- representation scheme
- deep learning
- feature selection
- image segmentation
- semi supervised
- feature representation
- scene recognition
- visual words
- receptive fields
- expectation maximization
- scene classification
- machine learning
- text classification
- feature extraction
- bag of features
- visual recognition tasks
- restricted boltzmann machine
- region segmentation
- low level features
- visual recognition
- natural scenes
- pattern recognition
- hierarchical structure
- text mining
- kernel function
- image processing
- linear combination
- natural images
- model selection