A Lightweight Riemannian Covariance Matrix Convolutional Network for PolSAR Image Classification.
Junfei ShiWei WangHaiyan JinMengmeng NieShanshan JiPublished in: IEEE Trans. Geosci. Remote. Sens. (2024)
Keyphrases
- lightweight
- covariance matrix
- image classification
- convolutional network
- convolutional neural networks
- covariance matrices
- riemannian manifolds
- log euclidean
- riemannian metric
- coarse to fine
- feature extraction
- principal component analysis
- feature extractor
- lie group
- positive definite
- image features
- image representation
- sample size
- gaussian mixture
- fisher information
- objective function
- estimation error
- affine invariant
- sparse coding
- sparse representation
- shape analysis
- geometrical interpretation
- multiresolution
- shape space
- facial landmarks
- wireless sensor networks