DL3DV Dataset

ScanNet

Cambridge Landmarks

ExPose: Reinforcing Video Generation Models for Extreme Pose Estimation