1 paper · 1 filter
Jiaxing Zhao, Qize Yang, Yixing Peng +8
In human-centric scenes, the ability to simultaneously understand visual and auditory information is crucial. While recent omni models can process multiple modalities, they general…