We believe that the criticality graph and zones can be used to register images that represent threshold defined objects. We are developing tree matching algorithms in this regard, so that we can map the criticality trees of images of the same volume taken via different modalities. If a subtree T1 of G1 maps to a subtree T2 of G2, the appropriate tensor transformation between the zones of T1 and T2 can be computed, as part of the overall registration transformation. Note that the criticality graphs are invariant under translation, rotation, scaling, and value translation.