A Weakly-Supervised Change Detection for Multispectral Earth Observation Imagery using a Long Short- Term Memory Classifier with a Virtual Training Data Neural Generator.
Ionut-Alexandru GîrlaVictor-Emil NeagoePublished in: COMM (2022)
Keyphrases
- multispectral
- change detection
- remote sensing
- weakly supervised
- training data
- remote sensing data
- satellite imagery
- remote sensing images
- high resolution
- satellite data
- remotely sensed
- remote sensing imagery
- land cover
- satellite images
- remotely sensed images
- hyperspectral
- training set
- multispectral images
- image processing
- image analysis
- object class
- superpixels
- remotely sensed data
- training examples
- hyperspectral data
- learning algorithm
- semi supervised
- data sets
- neural network
- supervised learning
- image data
- hyperspectral images
- data sources
- data mining
- super resolution
- object detection
- active learning
- pattern recognition
- feature selection