Page 246 - Machine Learning for Subsurface Characterization
P. 246

Deep neural network architectures Chapter  7 211


                                                                     Manifold of T2 in the three-dimensional latent space as learned by VAEc from the 460 depths in the training dataset. x-axis represents T2 relaxation time























                                                                       (unitless).


                                                                       0.25
                                                                       to
                                                                       0
                                                                       from
                                                                       ranging

                                                                       amplitude



                                                                       signal
                                                                       NMR

                                                                       normalized



                                                                       is
                                                                       y-axis
                                                                       3000 ms;


                                                                       to
                                                                     FIG. 7.13  0.3  from
   241   242   243   244   245   246   247   248   249   250   251