GraspClutter6D

From the catalogue. Listed by one researcher from one source page.

• id: graspclutter6d
• kind: grasping
• robot or device: four RGB-D cameras; bins, shelves and tables (class: sensor only)
• how collected: real capture, pose and grasp annotation
• size as stated: 1,000 scenes; 200 objects; 52K RGB-D images; 736K 6D poses; 9.3B grasps
• hours: not stated | hours only claimed: none | episodes: not stated
• year: 2025
• organisation: GIST, KAIST et al.
• licence: CC BY-SA 4.0 (class: open)
• access: open (class: open)
• link (dataset): https://huggingface.co/datasets/GraspClutter6D/GraspClutter6D
• Hugging Face repository: GraspClutter6D/GraspClutter6D, 1,128 downloads in 30 days, https://huggingface.co/datasets/GraspClutter6D/GraspClutter6D
• note: 14.1 objects per scene, 62.6 percent occlusion.

Confirm the licence at the link before relying on it.