|TREND||DATASET||BEST METHOD||PAPER TITLE||PAPER||CODE||COMPARE|
Particularly on small displacements and real-world data, FlowNet cannot compete with variational methods.
It then uses the warped features and features of the first image to construct a cost volume, which is processed by a CNN to estimate the optical flow.
Ranked #2 on Dense Pixel Correspondence Estimation on HPatches
We learn to compute optical flow by combining a classical spatial-pyramid formulation with deep learning.
Ranked #5 on Dense Pixel Correspondence Estimation on HPatches