bit-da/msra

Code release for "End-to-End Transferable Anomaly Detection via Multi-spectral Cross-domain Representation Alignment"

End-to-End Transferable Anomaly Detection via Multi-spectral Cross-domain Representation Alignment

Code release for "End-to-End Transferable Anomaly Detection via Multi-spectral Cross-domain Representation Alignment"

Paper

In this paper, we propose a Multi-spectral Cross-domain Representation Alignment (MsRA) method for the anomaly detection in the domain adaptation setting, where we can only access a set of normal source data and a limited number of normal target data.

Prerequisites

The code is implemented with CUDA 10.0.130, Python 3.6.13 and Pytorch 1.2.0.

To install the required python packages, run

pip install -r requirements.txt

Datasets

Download the dataset and place the images to the corresponding folder.

Office-Home dataset can be found here.

Running the code

python MsRA.py --dataset OfficeHomeDataset --source Product --target Clipart --c_cls Bike

Acknowledgements

Some codes are adapted from DSVDD, DANN and BiOST. We thank them for their excellent projects.

Contact

If you have any problem about our code, feel free to contact

• [email protected]
• [email protected]
• [email protected]
• [email protected]

or describe your problem in Issues.