Complex-Valued Convolutional Neural Networks in Interferometric Synthetic Aperture Radar and Their Teacher-Image Pollution Influence on the Performance.
Yuki SunagaRyo NatsuakiAkira HirosePublished in: IGARSS (2020)
Keyphrases
- sar imagery
- synthetic aperture radar
- sar images
- complex valued
- sar data
- sar imaging
- synthetic aperture
- radar imaging
- image reconstruction
- real valued
- space variant
- spectral estimation
- image features
- edge detection
- target detection
- high resolution
- multiscale
- convolutional neural networks
- inverse synthetic aperture radar
- image segmentation
- input image
- image retrieval
- activation function
- urban areas
- aerial images
- edge detector
- maximum likelihood
- image classification
- signal processing
- neural network