problems about the paper

Thanks for the great work!
I'd like to know what's the difference between the Sample Field S and Se, is it just different in the number of FC layers? Using the output of the Sample Field S to get MSE loss and using the output of Sample Field Se to do depth boost? And I'm also not very clear the 
physical meaning about the formulation of depth boost loss Ld.
Thanks.