pipette
ENEnglish

Anaximander: Interactively Running Geospatial Deep Learning Models on Any Compute Backend

Satej S. Soman, Akram Zaytar, Girmaw A. Tadesse, Gilles Q. Hacheme, Muhammad S. Danish, Inbal Becker-Reshef, Rahul Dodhia, Juan Lavista Ferres

PreprintUso en el mundo realCódigo disponible

En palabras de los autores

Applying deep learning models to satellite imagery from within geographic information systems (GIS) remains high-friction for remote sensing practitioners. Models arrive in incompatible formats and target different compute environments, from local workstations to serverless cloud services. As a result, every evaluation demands custom deployment, tiling, and georeferencing code before a single prediction reaches the analyst's map. This friction discourages systematic comparison in a domain where model choice directly affects operational outcomes such as field delineation, crop monitoring, and disaster response. We present Anaximander, an open-source system that unifies model source and compute location choice behind one interactive interface. The system's backend is an inference server that loads models from multiple commonly-used sources and serves them on any accessible compute backend. The server provides session management and model caching, and streams results back per tile. The backend is paired with a QGIS plugin that drives tiling, result reassembly, georeferencing, and real-time per-tile status visualization. An additional user-interface path injects layer legends as prompts into vision-language models. We demonstrate the system in a code-free side-by-side comparison of three heterogeneous models on an agricultural field delineation task: gpt-image-1 via a cloud API, Segment Anything Model 3 (SAM3) on a remote GPU, and DelineateAnything on a local CPU. The inference backend and protocol are open-source and available at https://github.com/microsoft/nxmndr.

Resultado principalEl resumen no menciona limitaciones.

Apareció: jueves, 8 de octubre. arXiv. Preprint, todavía sin revisión por pares.

Comentario de los autores: Accepted as a poster at the TerraBytes II workshop, ECCV 2026