The Gap Is in the Geometry: A Convergent Meta-Analysis of Multi-Facet Aerial and Remote Sensing Video Understanding

A new research paper analyzes multimodal large language models on aerial and remote sensing video benchmarks, revealing a large performance gap driven by…