3 accepted papers
Spatiotemporal fusion (STF) bridges the gap between temporal and spatial resolutions in satellite imagery, enabling effective monitoring of Earth's surface dynamics. However, existing methods rely on cloud-free reference images, a constraint that fails in realistic, cloud-prone scenarios. To overcom
Multimodal Large Language Models (MLLMs) have demonstrated exceptional proficiency in standard text extraction, but they encounter significant challenges when confronting real-world implicit text. Such content typically contains malicious information, intentionally concealed through physical deforma…
With the rapid advancement of generative models, general-purpose generation has gained increasing attention as a promising approach to unify diverse tasks across modalities within a single system. Despite this progress, existing open-source frameworks often remain fragile and struggle to support com…