Przemysław Głomb

Tool-using LVLM agents can significantly enhance hyperspectral unmixing by intelligently refining outputs from existing modular pipelines, leading to more accurate material decomposition.

Seasonal variations can drastically affect building detection accuracy, with specific months outperforming others in model training and inference.

Local LLMs can achieve cloud-level accuracy with prompt-side preprocessing, dramatically boosting performance while keeping latency low.