Data and code for weighted attention UNet for OBN/OBC denoising

This data and the associated codes were developed for the research article “Weighted Convolutional Attention UNet for Ocean Bottom Seismic Noise Suppression and Transfer Learning.”

The data include a semi-synthetic dataset and a 3D field ocean bottom node (OBN) test dataset.

In the semi-synthetic dataset, the directories train_data and val_data contain the files used for training and validation, respectively.

The data dimensions are nx × nz = 256 × 256.

The dataset files are named as follows:

Data containing ground roll and vertical stripe noise (input):

noisy_data

Ground roll and vertical stripe noise (output):

noise