1 paper · 1 filter
Jingli Lin, Runsen Xu, Shaohao Zhu +11
Spatial understanding over continuous visual input is crucial for MLLMs to evolve into general-purpose assistants in physical environments. Yet there is still no comprehensive benc…