Denoising Fault-Aware Wavelet Network: a nueral network constructed under the frame work of signal processing informed neural network (SPINN)
- parameter setting: main.py
- training program:utils/train_utils.py
- Spectral kurtosis based Loss: physics_loss/kurtosis.py
- Energy-based channel (scale) selection layer (attention weighting layer):paper_model/weighting_layer.py
- The whole DFAWnet:paper_model/DFWnet.py
- Fused 1/2/3 wavelet convolution:fusion_modules/fuse_conv_wavelet.py
- Dynamic hard thresholding mask:dynconv.py
@article{shang2023denoising,
title={Denoising Fault-Aware Wavelet Network: A Signal Processing Informed Neural Network for Fault Diagnosis},
author={Shang, Zuogang and Zhao, Zhibin and Yan, Ruqiang},
journal={Chinese Journal of Mechanical Engineering},
volume={36},
number={1},
pages={9},
year={2023},
publisher={Springer}
}