---
license: cc-by-4.0
language:
- en
pretty_name: Ego-Tactile Manipulation
size_categories:
- 100K<n<1M
task_categories:
- robotics
tags:
- LeRobot
- robotics
- physical-ai
- egocentric
- tactile
- manipulation
- action-segmentation
- multimodal
---

# Ego-Tactile Manipulation

*Egocentric video + dense two-hand tactile + touch-grounded action labels - by [**OpenGraph Labs**](https://www.opengraphlabs.com/).*

<video controls autoplay loop muted playsinline width="100%" src="https://huggingface.co/datasets/OpenGraphLabs-Research/ego-tactile-manipulation/resolve/main/assets/ego_tactile_grid.mp4"></video>

<sub>Four episodes visualized in our dashboard - egocentric video with live tactile & sensor signals.</sub>

Synchronized **ego + touch** human-manipulation data is rare. This is a clean **1.28-hour sample** from OpenGraph Labs' Physical-AI data pipeline: a head camera plus our **OGLO tactile gloves** on both hands, with action labels derived from the physical contact signal.

## OGLO - our tactile gloves

![OGLO tactile gloves](https://huggingface.co/datasets/OpenGraphLabs-Research/ego-tactile-manipulation/resolve/main/assets/oglo_gloves.png)

Worn on both hands · **5 fingers × 4×4 taxels = 80 channels per hand (160 total)** · 90 Hz pressure, co-registered to every video frame.

## Touch-grounded action segmentation

Action boundaries come from physical **contact and grip-force** - no human ground truth. Concise labels are added within each physically-fixed span.

![Action storyboard](https://huggingface.co/datasets/OpenGraphLabs-Research/ego-tactile-manipulation/resolve/main/assets/action_seg_storyboard.png)
![Action browser](https://huggingface.co/datasets/OpenGraphLabs-Research/ego-tactile-manipulation/resolve/main/assets/action_seg_browser.png)

## Specs

| | |
|---|---|
| Episodes / duration | 72 · **1.28 h** (138,629 frames @ 30 fps) |
| Camera | egocentric, 1920×1080 (intrinsics included) |
| Tactile | 160 channels · 30 Hz aligned + 90 Hz raw |
| IMU | head (9-ch) + both wrists (6-ch each) |
| Annotations | action / sub-action / anchor / grasp / hand / object (English) |
| Tasks | tabletop manipulation - pour, cap, cut, grip, arrange, … |
| License | CC-BY-4.0 |

## What's inside

```
meta/      LeRobot v3.0 metadata  (+ opengraph/: camera intrinsics, tactile highlights)
data/      per-frame parquet      (tactile @30 Hz, IMU, annotations)
videos/    egocentric video
extra/     raw_tactile/ + raw_imu/ (head, both wrists)   byte-exact raw sensors
rerun/     one .rrd per episode   (instant visualization)
```

## Format & use

Packaged in **[LeRobot v3.0](https://huggingface.co/docs/lerobot)** and **[Rerun](https://rerun.io)**, so it's ready to use as-is:

```python
from lerobot.datasets import LeRobotDataset
ds = LeRobotDataset("OpenGraphLabs-Research/ego-tactile-manipulation")
```
```bash
rerun rerun/episode_000000.rrd      # visualize an episode
```

## About OpenGraph Labs

Physical-AI data infrastructure - we capture and post-process large-scale egocentric human manipulation data (multi-camera, tactile, IMU, metric egomotion, frame-level semantics). This is our first public sample; reach out for collaborations and larger releases.

🌐 [opengraphlabs.com](https://www.opengraphlabs.com/) · 🤗 [OpenGraphLabs-Research](https://huggingface.co/OpenGraphLabs-Research)

## Citation

```bibtex
@misc{opengraphlabs2026egotactile,
  title  = {Ego-Tactile Manipulation},
  author = {OpenGraph Labs},
  year   = {2026},
  url    = {https://huggingface.co/datasets/OpenGraphLabs-Research/ego-tactile-manipulation}
}
```

Released under **CC-BY-4.0** - free to use, including commercially, with attribution.