Seismic data for target-adaptive low-frequency reconstruction

This study aims to address the limited generalization capability of traditional neural network-based low-frequency prediction methods by incorporating demigrated samples tailored to the target area for network training. The dataset includes a demigrated dataset for network training, a trained neural network, a synthetic high-frequency dataset simulating field band-limited conditions, and the low-frequency data predicted from the band-limited data using the trained network.