A Self-Supervised Hyperspectral Image Restoration Method Based on Convolutional Sparse Coding and Superpixel Segmentation.
Zhongshun JiangHonglin ZhuYi QiuYuntao QianPublished in: IGARSS (2022)
Keyphrases
- sparse coding
- restoration method
- hyperspectral images
- natural images
- sparse representation
- unsupervised learning
- linear combination
- image classification
- hyperspectral
- image restoration
- generative model
- image representation
- segmentation algorithm
- remote sensing
- image segmentation
- multiscale
- degraded images
- feature space
- image regions
- image patches
- segmentation method
- high dimensional
- region growing
- input image
- energy function
- test images
- higher order
- optical flow
- visual words
- feature vectors
- maximum likelihood