DiversePose 300

DiversePose 300 DiversePose 300 is a challenging category-level object pose estimation benchmark introduced in CPPF++: Uncertainty-Aware Sim2Real Object Pose Estimation by Vote Aggregation (TPAMI 2024). It contains real-world RGB-D captures of common objects (bottles, bowls, mugs) in diverse, unconstrained scenes, designed to test the generalization of pose estimators beyond NOCS REAL275. Code: Project page:… See the full description on the dataset page:

Source: Hugging Face Hub (qq456cvb/DiversePose300). Metadata imported from the dataset’s Hub tags.