Gradient-based deterministic inversion of geophysical data with Generative Adversarial Networks: is it feasible?
This is the companion code of the paper by Laloy et al. (see reference below). This Python 3 package contains the used pytorch-written GAN together with the tested two deterministic inversion strategies for the considered linear ground penetrating radar (GPR) tomography problem.
Laloy, E., Linde, N., Ruffino, C., Hérault, R., & Jacques, D. 2018. Gradient-based deterministic inversion of geophysical data with Generative Adversarial Networks: is it feasible? arXiv:1812.09140, https://arxiv.org/abs/1812.09140
MIT
Eric Laloy ([email protected])