Physics-Driven Learning of Wasserstein GAN for Density Reconstruction in Dynamic Tomography