Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,4 @@
|
|
|
|
|
| 1 |
---
|
| 2 |
dataset_info:
|
| 3 |
- config_name: full
|
|
@@ -44,104 +45,111 @@ configs:
|
|
| 44 |
default: true
|
| 45 |
---
|
| 46 |
|
|
|
|
| 47 |
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
<
|
| 51 |
-
|
| 52 |
-
## Dataset Summary
|
| 53 |
|
| 54 |
-
|
| 55 |
-
The
|
|
|
|
| 56 |
|
| 57 |
-
|
|
|
|
|
|
|
|
|
|
| 58 |
|
| 59 |
---
|
| 60 |
|
| 61 |
-
## Dataset
|
| 62 |
|
| 63 |
-
|
|
|
|
|
|
|
|
|
|
| 64 |
|
| 65 |
-
- **Actuated Angles**:
|
| 66 |
-
- 16 actuator values representing positions/angles controlling the face and eyes.
|
| 67 |
|
| 68 |
-
|
| 69 |
-
- RGB image of the humanoid face (resolution: **[FILL_RESOLUTION]**) matching the actuator positions.
|
| 70 |
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
|
| 79 |
-
|
|
| 80 |
-
|
|
| 81 |
-
| 01 | Cheek - Right | 09 | Eyelid Lower - Right |
|
| 82 |
-
| 02 | Eyeball Sideways - Left | 10 | Forehead - Right |
|
| 83 |
-
| 03 | Eyeball Up Down - Left | 11 | Forehead - Left |
|
| 84 |
-
| 04 | Eyelid Upper - Left | 12 | Upper Nose |
|
| 85 |
-
| 05 | Eyelid Lower - Left | 13 | Eyebrow - Right |
|
| 86 |
-
| 06 | Eyeball Up Down - Right | 14 | Jaw |
|
| 87 |
-
| 07 | Eyeball Sideways - Right | 15 | Eyebrow - Left |
|
| 88 |
|
| 89 |
---
|
| 90 |
|
|
|
|
| 91 |
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
**
|
| 95 |
|
| 96 |
-
<p align="center"> <img src="https://huggingface.co/datasets/iamirulofficial/Test2/resolve/main/Screenshot%202025-05-02%20at%204.04.28%E2%80%AFPM.png" alt="Full Face Actuator Map" width="35%"/> <img src="https://huggingface.co/datasets/iamirulofficial/Test2/resolve/main/Screenshot%202025-05-02%20at%204.03.56%E2%80%AFPM.png" alt="Eye Actuator Map" width="35%"/> </p>
|
| 97 |
|
| 98 |
---
|
| 99 |
|
| 100 |
-
## Dataset
|
| 101 |
|
| 102 |
-
|
| 103 |
-
-
|
|
|
|
|
|
|
| 104 |
|
| 105 |
-
|
| 106 |
|
| 107 |
-
|
| 108 |
|
| 109 |
-
|
| 110 |
|
| 111 |
```python
|
| 112 |
-
from datasets import load_dataset
|
| 113 |
|
| 114 |
-
#
|
| 115 |
-
ds = load_dataset("
|
|
|
|
| 116 |
|
| 117 |
-
|
| 118 |
-
ds
|
|
|
|
|
|
|
| 119 |
````
|
| 120 |
|
|
|
|
|
|
|
|
|
|
| 121 |
---
|
| 122 |
|
| 123 |
-
## Data
|
| 124 |
|
| 125 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 126 |
|
| 127 |
---
|
| 128 |
|
| 129 |
## License
|
| 130 |
|
| 131 |
-
|
| 132 |
|
| 133 |
---
|
| 134 |
|
| 135 |
## Citation
|
| 136 |
|
| 137 |
-
If you use this dataset, please cite:
|
| 138 |
-
|
| 139 |
```bibtex
|
| 140 |
-
@misc{
|
| 141 |
-
title
|
| 142 |
-
author
|
| 143 |
-
year
|
| 144 |
-
url
|
| 145 |
}
|
| 146 |
```
|
| 147 |
|
|
@@ -149,6 +157,5 @@ If you use this dataset, please cite:
|
|
| 149 |
|
| 150 |
## Acknowledgements
|
| 151 |
|
| 152 |
-
|
| 153 |
-
|
| 154 |
|
|
|
|
| 1 |
+
|
| 2 |
---
|
| 3 |
dataset_info:
|
| 4 |
- config_name: full
|
|
|
|
| 45 |
default: true
|
| 46 |
---
|
| 47 |
|
| 48 |
+
# 🧠 Open Humanoid Actuated Face Dataset
|
| 49 |
|
| 50 |
+
<p align="center">
|
| 51 |
+
<img src="https://huggingface.co/datasets/iamirulofficial/Test2/resolve/main/imgesFace.png" alt="Sample Face Image" width="360"/>
|
| 52 |
+
</p>
|
|
|
|
|
|
|
| 53 |
|
| 54 |
+
## Dataset Summary
|
| 55 |
+
The **Open Humanoid Actuated Face Dataset** is designed for researchers working on
|
| 56 |
+
facial‑actuation control, robotics, reinforcement learning, and human–computer interaction.
|
| 57 |
|
| 58 |
+
* **Origin** – collected during a reinforcement‑learning (RL) training loop whose objective was to reproduce human facial expressions.
|
| 59 |
+
* **Platform** – a modified **i2Head InMoov** humanoid head with a silicone skin.
|
| 60 |
+
* **Control** – **16 actuators** driving facial features and eyeballs.
|
| 61 |
+
* **Pairing** – each example contains the raw RGB image **and** the exact actuator angles that produced it.
|
| 62 |
|
| 63 |
---
|
| 64 |
|
| 65 |
+
## Dataset Structure
|
| 66 |
|
| 67 |
+
| Field | Type | Description |
|
| 68 |
+
|-----------------|---------------|---------------------------------------------------------------|
|
| 69 |
+
| `image` | `Image` | RGB capture of the humanoid face (resolution **[FILL_RES]**). |
|
| 70 |
+
| `actuated_angle`| `struct` | 16 integer values (`"0"`, `"1"` .. so on) |
|
| 71 |
|
|
|
|
|
|
|
| 72 |
|
| 73 |
+
### Actuator Index Reference
|
|
|
|
| 74 |
|
| 75 |
+
| Idx | Actuator | Idx | Actuator |
|
| 76 |
+
|:---:|--------------------------------|:---:|-------------------------|
|
| 77 |
+
| 00 | Cheek – Left | 08 | Eyelid Upper – Right |
|
| 78 |
+
| 01 | Cheek – Right | 09 | Eyelid Lower – Right |
|
| 79 |
+
| 02 | Eyeball Sideways – Left | 10 | Forehead – Right |
|
| 80 |
+
| 03 | Eyeball Up/Down – Left | 11 | Forehead – Left |
|
| 81 |
+
| 04 | Eyelid Upper – Left | 12 | Upper Nose |
|
| 82 |
+
| 05 | Eyelid Lower – Left | 13 | Eyebrow – Right |
|
| 83 |
+
| 06 | Eyeball Up/Down – Right | 14 | Jaw |
|
| 84 |
+
| 07 | Eyeball Sideways – Right | 15 | Eyebrow – Left |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
|
| 86 |
---
|
| 87 |
|
| 88 |
+
### Actuator Mapping Images *(placeholders)*
|
| 89 |
|
| 90 |
+
| Full‑Face Map | Eye‑Only Map |
|
| 91 |
+
|:-------------:|:-----------:|
|
| 92 |
+
| *![ADD IMAGE]*<br><img src="https://huggingface.co/datasets/iamirulofficial/Test2/resolve/main/Screenshot%202025-05-02%20at%204.04.28%E2%80%AFPM.png" width="50%"/> | *![ADD IMAGE]*<br><img src="https://huggingface.co/datasets/iamirulofficial/Test2/resolve/main/Screenshot%202025-05-02%20at%204.03.56%E2%80%AFPM.png" width="50%"/> |
|
| 93 |
|
|
|
|
| 94 |
|
| 95 |
---
|
| 96 |
|
| 97 |
+
## Dataset Statistics
|
| 98 |
|
| 99 |
+
| Split | Samples | Size |
|
| 100 |
+
|-------|---------|------|
|
| 101 |
+
| **Train (full)** | **[FILL_TOTAL]** | ≈ 105 GB |
|
| 102 |
+
| **Train (small)** | 2 | ≈ 2 GB |
|
| 103 |
|
| 104 |
+
*(Numbers shown are for the preview release — update as you add data.)*
|
| 105 |
|
| 106 |
+
---
|
| 107 |
|
| 108 |
+
## Usage Example
|
| 109 |
|
| 110 |
```python
|
| 111 |
+
from datasets import load_dataset, Image
|
| 112 |
|
| 113 |
+
# load the small subset
|
| 114 |
+
ds = load_dataset("iamirulofficial/Test2", name="small", split="train")
|
| 115 |
+
ds = ds.cast_column("image", Image()) # decode image bytes ➜ PIL.Image
|
| 116 |
|
| 117 |
+
img = ds[0]["image"]
|
| 118 |
+
angles = ds[0]["actuated_angle"] # {'0': 90, '1': 20, ...}
|
| 119 |
+
img.show()
|
| 120 |
+
print(angles)
|
| 121 |
````
|
| 122 |
|
| 123 |
+
> **Tip**
|
| 124 |
+
> For the full corpus use `name="full"` (may require `streaming=True` once the dataset grows).
|
| 125 |
+
|
| 126 |
---
|
| 127 |
|
| 128 |
+
## Data Collection & RL Setup
|
| 129 |
|
| 130 |
+
A detailed description of the RL pipeline, reward design, and actuator hardware will appear in our upcoming paper (in preparation, 2025). Briefly:
|
| 131 |
+
|
| 132 |
+
1. **Vision module** extracts target expression keypoints from live human video.
|
| 133 |
+
2. **Policy network** predicts 16 actuator set‑points.
|
| 134 |
+
3. **Real‑time reward** computes expression similarity + smoothness penalties.
|
| 135 |
+
4. Images & angle vectors are logged every *N* steps, forming this dataset.
|
| 136 |
|
| 137 |
---
|
| 138 |
|
| 139 |
## License
|
| 140 |
|
| 141 |
+
Released under the **MIT License** – free for commercial and non‑commercial use.
|
| 142 |
|
| 143 |
---
|
| 144 |
|
| 145 |
## Citation
|
| 146 |
|
|
|
|
|
|
|
| 147 |
```bibtex
|
| 148 |
+
@misc{amirul2025openhumanoidface,
|
| 149 |
+
title = {Open Humanoid Actuated Face Dataset},
|
| 150 |
+
author = {Imtiyaz Amirul et al.},
|
| 151 |
+
year = {2025},
|
| 152 |
+
url = {https://huggingface.co/datasets/iamirulofficial/Test2}
|
| 153 |
}
|
| 154 |
```
|
| 155 |
|
|
|
|
| 157 |
|
| 158 |
## Acknowledgements
|
| 159 |
|
| 160 |
+
Big thanks to the **i2Head InMoov** community and everyone who helped engineer the silicone skin and actuator stack.
|
|
|
|
| 161 |
|