Seismic denoising by deep learning from natural repeating earthquakes

We propose that similar waveforms from repeating earthquakes can be used to evaluate the performance of seismic denoising methods. Our training dataset is from natural repeating earthquakes, without any synthetic data as labels or assumptions about the noise. Once downloaded, you'll find a folder named "example". Inside this folder are sample SAC data and the denoising program.The provided script (written in Python) processes the SAC data to remove noise. Simply place your SAC files into the "example" folder, replacing the provided SAC example data.