1 paper · 1 filter
Changpeng Wang, Xin Lin, Junhan Liu +5
Multimodal large laboratory models (MLLMs) still struggle with spatial understanding under the dominant perspective-image paradigm, which inherits the narrow field of view of human…