TriO: Tri-Modal Unsupervised Occupancy World Model for Anything Perception
TriO is a multi-modal unsupervised world model that predicts 4D occupancy, obstacle segmentation, flow and LiDAR, and is able to segment any occupancy from the drivable surface, overcoming the limitations of existing open-set methods in handling long-tail objects.
Quinlan Sykora, Sourav Biswas, Christopher Diehl et al.
· 0 citations