One-instance fractals for anomaly detection

This repository implements the One-instance FractalDataBase (OFDB) [1], and apply it for defect detection. OFDB is a dataset of abstract fractal images, with only one sample per class.

We generate a dataset with 1,000 classes to train a ResNet-like backbones and leverage the pre-trained features for defect detection. Despite using only 1,000 samples, strong data augmentation allows the model to achieve results comparable to those pre-trained on ImageNet, which requires over a million samples and extensive training time.

Getting Started

The code uses PyTorch for training. Follow the official PyTorch installation guide to install the version compatible with your system. Additional required packages are listed in requirements.txt.

# clone project
git clone https://github.com/cugwu/OFDB-4AD.git
   
# [RECOMMENDED] set up a virtual environment
python -m venv venv_name  # choose your prefered venv_name
source venv_name/bin/activate

# install requirements
pip install -r requirements.txt

Dataset Generation

The dataset used for training is in ./1p-fractals. For the dataset generation we follow fractal4AD repository. The repository contains code for generating large-scale datasets of fractal images.

First an Iterated Function Systems (IFS), which serves as the basis for creating the final fractals dataset need to be generated

# generate a dataset of 100,000 systems, each with between 2 and 4 affine transformations
python ifs.py --save_path ./ifs-100k.pkl --num_systems 100000 --min_n 2 --max_n 4

This will produce a .pkl file containing a dictionary with the following structure:

{
  "params": [
    {"system": np.array(...)},
    ...
  ],
  "hparams": {
    ...
  }
}

The generated .pkl file can be used as an input parameter to generate the final fractals dataset. The following command generates a fractal dataset without a background, using the IFS dataset stored in ifs-100k.pkl. From this IFS database, 50,000 IFS samples will be used to create a dataset with 1,000 classes, each containing 1 sample of size 512x512:

python dataset_generator.py --dataset_type fractals --param_file ./ifs-100k.pkl --dataset_dir ./fractals_dataset --num_systems 50000 --num_class 1000 --per_class 1 --size 512

The dataset can also be generated directly without the .pkl file. In this case, the ifs.py script will be called internally by the dataset_generator.py script:

python dataset_generator.py --dataset_type fractals --dataset_dir ./fractals_dataset --num_systems 50000 --num_class 1000 --per_class 1 --size 512

Training

For training use the following commands:

python main.py --datadir ./1p-fractals --outdir ./fractals_pretrain  --store_name fractals_resnet50 --arch resnet50 --num_class 1000 --batch_size 128 --epochs 20000

We also offer additional backbone architectures: Fast Fourier Convolution [2] and Global Filter Networks [3]. Users can select their preferred model by setting the --arch parameter. The full list of available models can be found in networks.py.

Anomaly Detection Results

We provide the results on ResNet-50 and WideResNet-50 models pre-trained on ImageNet and OFDB-1k, as anomaly detection method we use the original code of PatchCore. In the table are reported respectively: instance_auroc, full_pixel_auroc, and anomaly_pixel_auroc.

MVTec PatchCore	resnet50	wide_resnet50
IN	0.989/0.981/0.973	0.991/0.981/0.973
OFDB-1k	0.937/0.965/0.956	0.936/0.959/0.949

License

This project is distributed under the Apache License, Version 2.0. See LICENSE.txt for more details. Some parts of the code are licensed under the MIT License. For those cases, the MIT License text is included at the top of the respective files.

Citation

If you use this code or find our work helpful for your resarch, please cite the following papers:

@misc{OFDB-4AD,
  author = {cugwu},
  title = {One-instance fractals for anomaly detection},
  year = {2025},
  url = {https://github.com/cugwu/OFDB-4AD},
  note = {Accessed: 2025-03-11}
}

References

[1] Nakamura, R., Kataoka, H., Takashima, S., Noriega, E. J. M., Yokota, R., & Inoue, N. (2023). Pre-training vision transformers with very limited synthesized images. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 20360-20369).

[2] Chi, L., Jiang, B., & Mu, Y. (2020). Fast fourier convolution. Advances in Neural Information Processing Systems, 33, 4479-4488.

[3] Rao, Y., Zhao, W., Zhu, Z., Zhou, J., & Lu, J. (2023). GFNet: Global filter networks for visual recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(9), 10960-10973.