Can Scale-Consistent Monocular Depth Be Learned in a Self-Supervised Scale-Invariant Manner? | IEEE Conference Publication | IEEE Xplore