README.md

---
license: cc-by-4.0
---
# Dataset Card for MPI3D-complex

## Dataset Description

The **MPI3D-complex dataset** is a **real-world image dataset** of **complex everyday objects**, designed for benchmarking algorithms in **disentangled representation learning** and **robustness to object variability**. It is an **extension** of the broader MPI3D dataset suite, which also includes [synthetic toy](https://huggingface.co/datasets/randall-lab/mpi3d-toy), [realistic simulated](https://huggingface.co/datasets/randall-lab/mpi3d-realistic), and [real-world geometric shape](https://huggingface.co/datasets/randall-lab/mpi3d-real) variants.

The **complex version** was recorded using the same **robotic platform** as the other MPI3D datasets, but with a new set of **real-world objects** (coffee-cup, tennis-ball, croissant, beer-cup) and a reduced color set. Images were captured using **real cameras**, with realistic lighting and background conditions. This allows researchers to assess how well models trained on simpler domains generalize to more complex object appearances.

All images depict **real-world objects** under **controlled variations of 7 known factors**:

- Object color (4 values)
- Object shape (4 values)
- Object size (2 values)
- Camera height (3 values)
- Background color (3 values)
- Robotic arm horizontal axis (40 values)
- Robotic arm vertical axis (40 values)

The dataset contains **460,800 images** at a resolution of **64×64 pixels**. The factors of variation are consistent with the other MPI3D datasets, but with **different factor sizes** for object shape and color.
![Dataset Visualization](https://huggingface.co/datasets/randall-lab/mpi3d-complex/resolve/main/complex1.gif)

## Dataset Source
- **Homepage**: [https://github.com/rr-learning/disentanglement_dataset](https://github.com/rr-learning/disentanglement_dataset)
- **License**: [Creative Commons Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)
- **Paper**: Muhammad Waleed Gondal et al. _On the Transfer of Inductive Bias from Simulation to the Real World: A New Disentanglement Dataset_. NeurIPS 2019.  

## Dataset Structure
|Factors|Possible Values|
|---|---|
|object_color|yellow=0, green=1, olive=2, red=3|
|object_shape|coffee-cup=0, tennis-ball=1, croissant=2, beer-cup=3|
|object_size|small=0, large=1|
|camera_height|top=0, center=1, bottom=2|
|background_color|purple=0, sea green=1, salmon=2|
|horizontal_axis (DOF1)|0,...,39|
|vertical_axis (DOF2)|0,...,39|

Each image corresponds to a unique combination of these 7 factors. The images are stored in a **row-major order** (fastest-changing factor is `vertical_axis`, slowest-changing factor is `object_color`).

### Why no train/test split?
The MPI3D-complex dataset does not provide an official train/test split. It is designed for **representation learning research** and for testing **robustness to object complexity and appearance variation**. Since the dataset is a complete Cartesian product of all factor combinations, models typically require access to the full dataset to explore factor-wise variations.

## Example Usage
Below is a quick example of how to load this dataset via the Hugging Face Datasets library:
```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("randall-lab/mpi3d-complex", split="train", trust_remote_code=True)
# Access a sample from the dataset
example = dataset[0]
image = example["image"]
label = example["label"]  # [object_color: 0, object_shape: 0, object_size: 0, camera_height: 0, background_color: 0, horizontal_axis: 0, vertical_axis: 0]
color = example["color"] # 0 
shape = example["shape"] # 0
size = example["size"] # 0
height = example["height"] # 0
background = example["background"] # 0 
dof1 = example["dof1"] # 0 
dof2 = example["dof2"] # 0

image.show()  # Display the image
print(f"Label (factors): {label}")
```
If you are using colab, you should update datasets to avoid errors
```
pip install -U datasets
```
## Citation
```
@article{gondal2019transfer,
  title={On the transfer of inductive bias from simulation to the real world: a new disentanglement dataset},
  author={Gondal, Muhammad Waleed and Wuthrich, Manuel and Miladinovic, Djordje and Locatello, Francesco and Breidt, Martin and Volchkov, Valentin and Akpo, Joel and Bachem, Olivier and Sch{\"o}lkopf, Bernhard and Bauer, Stefan},
  journal={Advances in Neural Information Processing Systems},
  volume={32},
  year={2019}
}
```