{"ID":2823999,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.24172","arxiv_id":"2512.24172","title":"Deep Global Clustering for Hyperspectral Image Segmentation: Concepts, Applications, and Open Challenges","abstract":"Hyperspectral imaging (HSI) analysis faces computational bottlenecks due to massive data volumes that exceed available memory. While foundation models pre-trained on large remote sensing datasets show promise, their learned representations often fail to transfer to domain-specific applications like close-range agricultural monitoring where spectral signatures, spatial scales, and semantic targets differ fundamentally. This report presents Deep Global Clustering (DGC), a conceptual framework for memory-efficient HSI segmentation that learns global clustering structure from local patch observations without pre-training. DGC operates on small patches with overlapping regions to enforce consistency, enabling training in under 30 minutes on consumer hardware while maintaining constant memory usage. On a leaf disease dataset, DGC achieves background-tissue separation (mean IoU 0.925) and demonstrates unsupervised disease detection through navigable semantic granularity. However, the framework suffers from optimization instability rooted in multi-objective loss balancing: meaningful representations emerge rapidly but degrade due to cluster over-merging in feature space. We position this work as intellectual scaffolding - the design philosophy has merit, but stable implementation requires principled approaches to dynamic loss balancing. Code and data are available at https://github.com/b05611038/HSI_global_clustering.","short_abstract":"Hyperspectral imaging (HSI) analysis faces computational bottlenecks due to massive data volumes that exceed available memory. While foundation models pre-trained on large remote sensing datasets show promise, their learned representations often fail to transfer to domain-specific applications like close-range agricult...","url_abs":"https://arxiv.org/abs/2512.24172","url_pdf":"https://arxiv.org/pdf/2512.24172v1","authors":"[\"Yu-Tang Chang\",\"Pin-Wei Chen\",\"Shih-Fang Chen\"]","published":"2025-12-30T12:10:43Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.LG\"]","methods":"[]","has_code":false,"code_links":[{"ID":605539,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2823999,"paper_url":"https://arxiv.org/abs/2512.24172","paper_title":"Deep Global Clustering for Hyperspectral Image Segmentation: Concepts, Applications, and Open Challenges","repo_url":"https://github.com/b05611038/HSI_global_clustering","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}