Deep Learning and Superpixel Feature Extraction Based on Contractive Autoencoder for Change Detection in SAR Images.
Ning LvChen ChenTie QiuArun Kumar SangaiahPublished in: IEEE Trans. Ind. Informatics (2018)
Keyphrases
- deep learning
- change detection
- sar images
- feature extraction
- restricted boltzmann machine
- gamma distributions
- synthetic aperture radar
- remote sensing
- superpixels
- weakly supervised
- unsupervised learning
- remote sensing images
- machine learning
- image processing
- wavelet domain
- feature vectors
- data streams
- multiscale
- face recognition
- automatic target recognition
- image registration
- sar imagery
- satellite images
- frequency domain
- mental models
- dimensionality reduction
- wavelet transform
- feature selection
- pattern recognition
- edge detector
- image classification
- image analysis
- feature space
- computer vision
- infrared
- hyperspectral
- object segmentation
- maximum likelihood
- denoising