Robotics · Grasping · EGAD!Benchmark detail
Grasp benchmark · Parallel-jaw

EGAD!

parallel-jawdiagnosticevolved objects3D-printable

EGAD! (the Evolved Grasping Analysis Dataset) uses evolutionary algorithms to generate a set of objects that spans the space of geometric complexity and grasp difficulty, plus a 49-object 3D-printable evaluation set.

It is a diagnostic tool rather than a leaderboard: it tells you where a grasping method fails along the geometry/difficulty axes, not a single headline score.

Primary source
At a glance
Source
Morrison et al., RA-L 2020
Year
2020
Scale
2,000+ evolved objects · 49 diverse 3D-printable eval objects
Gripper
Parallel-jaw
Modality
Mesh · depth
Best-known
Diagnostic set (geometry × difficulty) · no single SOTA number
Key results
  • Objects span geometric complexity × grasp difficulty
  • 49 reproducible, 3D-printable evaluation objects
  • Diagnostic — no single SOTA number by design

SIM = simulation result · HW = physical hardware. Image-wise accuracy is detection quality, not real-robot pick success. Figures cited from Morrison et al., RA-L 2020.

Related benchmarks

← Back to the grasping register
Parallel-jaw

GraspNet-1Billion

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

Parallel-jaw

Dex-Net 2.0

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

Parallel-jaw

Grasp-Anything

Language-driven grasp synthesis · open-vocabulary scenes

Parallel-jaw

Jacquard

~95% image-wise (GR-ConvNet-class)