Dataset for Inversion of GISAXS data (1 layer)

The dataset consists of 50000 X-ray diffraction patterns simulated by BornAgain [1] software. For each simulation, a multilayer sample model of the following structure was used: air, tantalum oxide, silicon dioxide, and substrate. Parameters of each layer but tantalum oxide were kept fixed. Hence, each diffraction pattern depends on the parameters of the tantalum oxide layer: real and complex part of refractive index, thickness, roughness, Hurst parameter, and correlation length. Each simulation output is stored in an h5py file consisting of 1) diffraction image as a NumPy array of shape [1024, 512]; 2) sample parameters as a NumPy array with 6 elements. For further details regarding simulation, see https://github.com/maxxxzdn/gisaxs-reconstruction/simulation/simulation.

[1] Pospelov, G., Van Herck, W., Burle, J., Carmona Loaiza, J.M., Durniak, C., Fisher, J., Ganeva, M., Yurov, D., & Wuttke, J. (2020). BornAgain: software for simulating and fitting grazing-incidence small-angle scattering. Journal of Applied Crystallography, 53, 262 - 276.