Dataset

HA-Ego-Samples

A large-scale human sensorimotor effort collecting multimodal first-person recordings of real human activity.

open sourcebattle readyModel hawor-cvpr-2025
Source dataset reference
Videos
1K
Steps
1K
Evaluation packs
4
Technical quality
84.0
Annotation quality
85.0
Hand signal
8.0
Motion density
72.4

Ranked Categories

Public scorecards

Overall QualityProvisional

82.6

945 clips · 4419 votes · 27 raters

Hand-Object Interaction
82.6
Hand-Object InteractionProvisional

82.6

945 clips · 4419 votes · 27 raters

Reviewed clean confidence
86.7%
Reviewed clip coverage
94.5%
Confirmed issue rate
11.9%
Confirmed clean clips
826
Confirmed issue clips
119
Needs more votes
55
Verdicts
4419

Diagnostics

Non-ranked dataset signals

TrackabilityReady

84.5

200 clips · model hawor-cvpr-2025

Lost-track rate
10.0%
Relocalizations / min
1.4
IMU-only fraction
10.0%
Median feature count
129
Static pose jitter
0.50 cm
HygieneReady

82.7

200 clips · model hawor-cvpr-2025

Blur score
80.0%
Exposure score
80.0%
Dropped frame rate
0.0%
Sync offset
29 ms
Calibration completeness
80.0%
Hand SignalReady

27.7

200 clips · model hawor-cvpr-2025

Visible-hand fraction
20.0%
Mean joint confidence
30.0%
Hand temporal stability
40.0%
Interaction DensityReady

36.0

200 clips · model hawor-cvpr-2025

Hand-object co-presence
40.0%
Near-contact fraction
30.0%
Coverage / DiversityReady

57.5

200 clips · model hawor-cvpr-2025

Task-tag coverage
50.0%
Asset-title diversity
0.0%
Video coverage
100.0%
Step coverage
100.0%

Evaluation Packs

Dataset drilldowns