Robotics · Grasping · YCB Object & Model SetBenchmark detail
Grasp benchmark · Object set

YCB Object & Model Set

object setreproducibilityinfrastructurephysical

The YCB (Yale–CMU–Berkeley) Object and Model Set is a standardized set of 77 everyday physical objects with high-quality RGB-D scans and meshes, designed so manipulation results in different labs are reproducible on the same objects.

It is infrastructure, not a scored benchmark — but it underpins countless grasping and manipulation evaluations and the YCB-Video 6-DoF pose dataset.

Primary source
At a glance
Source
Calli et al., IEEE R&A Magazine 2015
Year
2015
Scale
77 physical objects + RGB-D scans & meshes
Gripper
Object set
Modality
RGB-D meshes
Best-known
Standard physical object set — not a scored benchmark
Key results
  • 77 physical objects with calibrated RGB-D scans + meshes
  • Enables cross-lab reproducibility of manipulation results
  • Basis for the YCB-Video pose benchmark

SIM = simulation result · HW = physical hardware. Image-wise accuracy is detection quality, not real-robot pick success. Figures cited from Calli et al., IEEE R&A Magazine 2015.

Related benchmarks

← Back to the grasping register
Parallel-jaw

GraspNet-1Billion

De-facto clutter benchmark · AnyGrasp current SOTA (AP)

Ambidextrous (jaw + suction)

Dex-Net 4.0

HW: 95% reliability · 300 MPPH (ABB YuMi)

Suction

SuctionNet-1Billion

HW: 80.65% grasp success · 100% object clearance (their method)

Parallel-jaw

Dex-Net 2.0

HW: 93% on adversarial · 99% precision on 40 novel objects (YuMi)