Dataset for A physics-guided neural network-based inversion framework for monitoring of CO2 storage via angle-dependent seismic images: NumPy-Based Simulation Data

This dataset contains 4,800 NumPy (.npy) files generated for the experiments presented in the associated EAGE conference paper. The data are organized into 2 subsets, saturation_maps.zip contains the groundtruth and prediction results, and synthetic_angle_gathers.zip contains angle-dependent seismic images which are input data for training. The dataset supports all preprocessing, training, and evaluation workflows provided in the accompanying GitHub repository. Due to size constraints, only a small representative subset is hosted on GitHub, while the complete dataset is archived here for reproducibility and transparency.