{"paper":{"title":"SpectralTrain: A Universal Framework for Hyperspectral Image Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"SpectralTrain speeds hyperspectral image model training by 2-7x using curriculum learning and PCA spectral reduction while keeping accuracy close to full training.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jiarui Zhao, Liping Yu, Meihua Zhou, Nan Wan, Ruiguo Hu, Wai Kin Fung, Wenzhuo Liu, Xinyu Tong","submitted_at":"2025-11-20T06:19:26Z","abstract_excerpt":"Hyperspectral image (HSI) classification typically involves large-scale data and computationally intensive training, which limits the practical deployment of deep learning models in real-world remote sensing tasks. This study introduces SpectralTrain, a universal, architecture-agnostic training framework that enhances learning efficiency by integrating curriculum learning (CL) with principal component analysis (PCA)-based spectral downsampling. By gradually introducing spectral complexity while preserving essential information, SpectralTrain enables efficient learning of spectral -- spatial pa"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SpectralTrain enables efficient learning of spectral-spatial patterns at significantly reduced computational costs, delivering consistent 2-7x training speedups with small-to-moderate accuracy deltas across backbones on three benchmark datasets.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That PCA-based spectral downsampling combined with a curriculum schedule preserves essential information for accurate classification while the gradual complexity increase reliably improves learning efficiency over standard training.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SpectralTrain is a universal training framework that combines curriculum learning and PCA spectral downsampling to deliver 2-7x faster training for hyperspectral image classification across multiple backbones and datasets with only small accuracy trade-offs.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"SpectralTrain speeds hyperspectral image model training by 2-7x using curriculum learning and PCA spectral reduction while keeping accuracy close to full training.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2395eaf2bd3882fb8d58b84ec623bb455df2f9a2097cfd23640867cea4a64602"},"source":{"id":"2511.16084","kind":"arxiv","version":2},"verdict":{"id":"296e128e-55d8-4ce0-83ca-f7695384dfa9","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T20:54:41.007292Z","strongest_claim":"SpectralTrain enables efficient learning of spectral-spatial patterns at significantly reduced computational costs, delivering consistent 2-7x training speedups with small-to-moderate accuracy deltas across backbones on three benchmark datasets.","one_line_summary":"SpectralTrain is a universal training framework that combines curriculum learning and PCA spectral downsampling to deliver 2-7x faster training for hyperspectral image classification across multiple backbones and datasets with only small accuracy trade-offs.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That PCA-based spectral downsampling combined with a curriculum schedule preserves essential information for accurate classification while the gradual complexity increase reliably improves learning efficiency over standard training.","pith_extraction_headline":"SpectralTrain speeds hyperspectral image model training by 2-7x using curriculum learning and PCA spectral reduction while keeping accuracy close to full training."},"references":{"count":48,"sample":[{"doi":"","year":2025,"title":"IEEE Transactions on Geoscience and Remote Sensing (2025)","work_id":"dec1a8b6-ac40-4bb3-a461-02dd57566c29","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"El-Gabri, A.R., Aly, H.A., Ghoniemy, T.S.,et al.: DLRA-Net: Deep local residual attention network with contextual refinement for spectral super- resolution. Int. J. Comput. Vis.133, 1499–1531 (2025) h","work_id":"da4cc966-6280-4ae9-83c2-a1d19ac85a18","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1016/j.rse.2007.07.028","year":2009,"title":"Recent advances in techniques for hyperspectral image processing","work_id":"bd6ef533-3819-4a2b-aaaa-47d7acfaacea","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2002,"title":"IEEE Signal processing magazine19(1), 17–28 (2002)","work_id":"4986113e-7b93-4f6a-ac2c-6fd824cb41bc","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"ISPRS Journal of Photogrammetry and Remote Sensing158, 279–317 (2019) 29","work_id":"cc14eb11-19fd-4888-93bb-e9bf5568cc48","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":48,"snapshot_sha256":"66dba3eec6d770b906da771d080d8aa7cccbe5407c99703b4b0cf516301f1384","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}