Search is not available for this dataset
image
imagewidth (px) 1.28k
1.28k
| actuated_angle
dict |
|---|---|
{
"0": 90,
"1": 20
}
|
|
{
"0": 90,
"1": 0
}
|
🧠 Open Humanoid Actuated Face Dataset
Dataset Summary
The Open Humanoid Actuated Face Dataset is designed for researchers working on
facial‑actuation control, robotics, reinforcement learning, and human–computer interaction.
- Origin – collected during a reinforcement‑learning (RL) training loop whose objective was to reproduce human facial expressions.
- Platform – a modified i2Head InMoov humanoid head with a silicone skin.
- Control – 16 actuators driving facial features and eyeballs.
- Pairing – each example contains the raw RGB image and the exact actuator angles that produced it.
Dataset Structure
| Field | Type | Description |
|---|---|---|
image |
Image |
RGB capture of the humanoid face (resolution [FILL_RES]). |
actuated_angle |
struct |
16 integer values ("0", "1" .. so on) |
Actuator Index Reference
| Idx | Actuator | Idx | Actuator |
|---|---|---|---|
| 00 | Cheek – Left | 08 | Eyelid Upper – Right |
| 01 | Cheek – Right | 09 | Eyelid Lower – Right |
| 02 | Eyeball Sideways – Left | 10 | Forehead – Right |
| 03 | Eyeball Up/Down – Left | 11 | Forehead – Left |
| 04 | Eyelid Upper – Left | 12 | Upper Nose |
| 05 | Eyelid Lower – Left | 13 | Eyebrow – Right |
| 06 | Eyeball Up/Down – Right | 14 | Jaw |
| 07 | Eyeball Sideways – Right | 15 | Eyebrow – Left |
Actuator Mapping Images (placeholders)
| Full‑Face Map | Eye‑Only Map |
|---|---|
![]() |
![]() |
Dataset Statistics
| Split | Samples | Size |
|---|---|---|
| Train (full) | [FILL_TOTAL] | ≈ 105 GB |
| Train (small) | 2 | ≈ 2 GB |
(Numbers shown are for the preview release — update as you add data.)
Usage Example
from datasets import load_dataset, Image
# load the small subset
ds = load_dataset("iamirulofficial/OpenHumnoidDataset", name="small", split="train")
ds = ds.cast_column("image", Image()) # decode image bytes ➜ PIL.Image
img = ds[0]["image"]
angles = ds[0]["actuated_angle"] # {'0': 90, '1': 20, ...}
img.show()
print(angles)
Tip For the full corpus use
name="full"(may requirestreaming=Trueonce the dataset grows).
Data Collection & RL Setup
A detailed description of the RL pipeline, reward design, and actuator hardware will appear in our upcoming paper (in preparation, 2025). Briefly:
- Vision module extracts target expression keypoints from live human video.
- Policy network predicts 16 actuator set‑points.
- Real‑time reward computes expression similarity + smoothness penalties.
- Images & angle vectors are logged every N steps, forming this dataset.
License
Released under the MIT License – free for commercial and non‑commercial use.
Citation
@misc{amirul2025openhumanoidface,
title = {Open Humanoid Actuated Face Dataset},
author = {Amirul et al.},
year = {2025},
url = {https://huggingface.co/datasets/iamirulofficial/OpenHumnoidDataset}
}
Acknowledgements
Big thanks to the i2Head InMoov community and everyone who helped engineer the silicone skin and actuator stack.
- Downloads last month
- 5

