Registration techniques often treat the anatomy as a continuum â a single smoothly deforming elastic body. However, the imaging field-of-view often contains several anatomical structures, e.g. different organs and sub-parts thereof, not all of which may necessarily be attached to their surrounding via connective tissue. In contrast to being smoothed by standard techniques, sliding along those anatomical interfaces during motion can be resolved using state-of-the-art registration techniques utilizing total-variation regularization. In this paper, such sliding-aware registration method is studied to observe visceral sliding in MRI sequences during breathing. From automatic registration outputs, quantified sliding images are generated and validated, showing significant correlation with manual landmarks along clearly identifiable sliding interfaces. These sliding images can be useful in verifying biomechanical continuum assumptions, in determining radiotherapy treatment margins, and as a smoothness prior for subsequent segmentation.