Yue Zhou, Shujun Zhao, Xue Yang, Ruigang Li, Tianwen Zhang, Mengcheng Lan, Chaofeng Chen, Lingfei Ma · 2026 · arXiv
DOI: 10.48550/arXiv.2610.31304 · open-access PDF
Durante la era del aprendizaje profundo, las innovaciones en los modelos de visión de sensores remotos (RS) se centraron principalmente en la optimización de las arquitecturas de red para tareas específicas y la realización de capacitaciones end-to-end. Sin embargo, con la llegada de la era de los modelos de lenguaje de visión (VLMs), restringidos por las demandas sustanciales de recursos computacionales, el paradigma de investigación principal se ha desplazado hacia la adaptación de VLMs de propósito general utilizando datos de dominio vertical de RS para el ajuste fino, en lugar de modificar sus arquitecturas centrales.
Machine translation (Meta M2M100); the English original follows.
During the deep learning era, innovations in remote sensing (RS) vision models primarily focused on optimizing network architectures for specific tasks and conducting end-to-end training. However, with the advent of the Vision-Language Models (VLMs) era, constrained by substantial computational resource demands, the mainstream research paradigm has shifted towards adapting general-purpose VLMs using RS vertical domain data for fine-tuning, rather than modifying their core architectures. This transition marks a change in the innovation pathway from model-centric to data-centric. Despite the inc
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31304, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange