Search papers, labs, and topics across Lattice.
3
0
3
Fine-tuning vision-language models in pixel space leads to a dramatic boost in navigation performance, outperforming traditional methods reliant on 3D information.
Realistic pedestrian simulations can significantly improve the training and evaluation of human-aware navigation systems, bridging the gap between simulated and real-world interactions.
Current single-view mesh reconstruction methods falter under robot camera rotations, leading to critical errors in spatial reasoning.
