Article · IEEE Geoscience and Remote Sensing Letters · January 1, 2026
A Spiking and Memory-Enhanced State-Space Model for Hyperspectral Image Classification
Faiq Ahmad, Muhammad Usama, Usman Ghous, Danish Shehzad, Manuel Mazzara, Muhammad Imran Ahmad
Abstract
Hyperspectral imaging (HSI) captures hundreds of contiguous spectral bands and enables fine-grained material discrimination. We present SpikeHopMamba, a hybrid architecture that combines neuromorphic spike encoding, a linearized Hopfield energy module, and selective state-space (SSSM) blocks to produce compact, multi-modal patch tokens for HSI classification. SpikeHopMamba fuses static spectral tokens, sequentially encoded spiking features, and patch-level associative-energy cues, then models long-range spatial dependencies via efficient SSSM blocks. We train the model end-to-end with cross-entropy (and label smoothing) and report results averaged over five independent runs. Experiments on Hanchuan (HC), Salinas (SA), and OHID1–9 show that SpikeHopMamba attains competitive accuracy while preserving computational efficiency (OA up to 99.60%, κ up to 99.53%).
More by Manuel Mazzara
- Anchored Rank-Adversarial Fine-Tuning Pretrained Visual Place Recognition ModelsarXiv · 2026
- See the Future, Classify the Present: Stochastic Latent World Modeling for Hyperspectral ImagesarXiv · 2026
- Autoresearch with Coding Agents: Generalizers and Metric-Maximizers on Quran Recitation DataLecture notes in computer science · 2026 · Citations: 1
- What Counts as a Mistake? Annotating Recitation Events in Quran Memorization TranscriptsarXiv · 2026