code for: "Anomaly Detection in Airbus Helicopters with Healthy Accelerometer Data "

These codes address the problem of anomaly detection based solely on one-class classification. The approach converts high-dimensional vibration data into wavelet images, which are then analyzed using an autoencoder. The output is forwarded to small-scale algorithms for further processing.The dataset used is available at https://doi.org/10.3929/ethz-b-000415151, along with information related to the original paper. Please first download the dataset and place it in the same directory as the code repository before using it. Start running the code files in alphabetical order.

The paper describing this work has been accepted for presentation at the ICCAD 2025 conference. The code will be open-sourced on the day the paper is presented and published in IEEE Xplore. More details, along with links to access both the paper and the code, will be made available here.