Skip to main content
Fig. 12 | Earth, Planets and Space

Fig. 12

From: Using TNT-NN to unlock the fast full spatial inversion of large magnetic microscopy data sets

Fig. 12

Horizontal transects through the center of the synthetic UMN logo results for scenario 3. Shown at left are the TNT-NN solution and known solution and at right is the difference thereof. Without additional regularization, least-squares methods act as low-pass filters which produce solutions that are naturally smooth. This can manifest as a “spreading” of the solution in space as seen at the right. This behavior does not allow the least-squares solution to fully capture the sharp interfaces in the known solution

Back to article page