Problems about 'compute_transformed_volume' function

[GuitarZH]

Thanks for your excellent work and well-done code.
I have read your paper in detail. However, I have a problem with 'compute_transformed_volume' function in geometry_until.py.
In Line 132~133, what do these codes mean? Can you explain these variables?
T_img_in = torch.matmul(T_gs_s, torch.matmul(K_s, torch.matmul(T_inv_list, K_d_inv)))
grid_img_in= torch.matmul(T_img_in, grid).squeeze(-1)

In addition, will you release the training code?

Thanks for your helps!