This repository contains an unofficial implementation of the paper: "DiffusionAD: Denoising Diffusion for Anomaly Detection" by Zhang, H., Wang, Z., Wu, Z., & Jiang, Y. G. (2023), which can be accessed here.
DiffusionAD introduces a novel technique to improve the anomaly detection task by applying denoising diffusion models. By employing this method, the model can discern between normal and abnormal data effectively, with improved performance on various datasets. This project replicates the research paper's methodology using Python.
Not Completed!!
To run this project, you'll need the following packages:
- Python 3.x
- PyTorch
- NumPy
- Matplotlib
- Scikit-learn
- CUDA for GPU acceleration (optional)
You can install the required packages using the following command:
shellCopy code
pip install -r requirements.txt
Note: Please ensure to use the appropriate pip command according to your Python environment setup (it could be pip, pip3, or python -m pip)
To run the main script:
shellCopy code
python main.py
You can adjust the model parameters in the config.py file to fine-tune the model according to your needs.
The original paper tests the model on multiple datasets. In our project, we have included code to handle the following datasets: (Name the datasets you've implemented here)
Please refer to the datasets/ directory for more details.
main.py: The main script to train and evaluate the model.model/: Contains the PyTorch implementation of the DiffusionAD model.datasets/: Contains the scripts to load and preprocess the datasets.config.py: Contains the configuration parameters for the model.requirements.txt: Contains the required packages to run this project.
(Name the metrics you're evaluating on here, such as ROC-AUC, Precision@k, etc.) on various datasets are as follows:
| Dataset | Metric1 | Metric2 | ... |
|---|---|---|---|
| Dataset1 | x.xx | x.xx | ... |
| Dataset2 | x.xx | x.xx | ... |
| ... | ... | ... | ... |
(Note: You might want to add plots/figures that highlight the results if any)
This project is licensed under the MIT License - see the LICENSE file for details.
This repository is an unofficial implementation of the paper: "DiffusionAD: Denoising Diffusion for Anomaly Detection". We would like to express our gratitude to the authors for their invaluable research and contribution to this field.
Zhang, H., Wang, Z., Wu, Z., & Jiang, Y. G. (2023). DiffusionAD: Denoising Diffusion for Anomaly Detection. arXiv preprint arXiv:2303.08730.
The Perlin noise code is adopted from VitjanZ/DRAEM.
This project is for educational purposes only. No responsibility is taken for any outcomes that arise from using this code.
Note: You can modify the sections as per your project's requirements. Remember to replace placeholders with actual values.