RIHP-Net: Rotation-Invariant Hyperspectral Image Classification via Harmonic Fiber Pooling and Rotation-Conditioned Spectral Fusion

Hyperspectral image classification can degrade under in-plane rotations because spatial structures change orientation and the learned spectral–spatial coupling in patch-based models becomes less stable. Most existing approaches improve robustness mainly through data augmentation, without expl...

Full beskrivning

Bibliografiska uppgifter

Huvudupphov:	Arafat Al-Jawari, Li Fang, Metwalli Al-Selwi, Huifang Shen
Materialtyp:	Artigo
Språk:	Inglês
Utgiven:	IEEE 2026-01-01
Serie:	IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Ämnen:	Group-equivariant convolution harmonic fiber pooling hyperspectral image classification (HSIC) rotation invariance rotation-conditioned spectral fusion (RCSF) spectral–spatial feature learning
Länkar:	https://ieeexplore.ieee.org/document/11534556/
Taggar:	Lägg till en tagg Inga taggar, Lägg till första taggen!