Submitted by Aljaz Bozic.

Submission data

Full nameFlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks
DescriptionIn this paper, we advance the concept of end-to-end learning of optical flow and make it work really well. The large improvements in quality and speed are caused by three major contributions: first, we focus on the training data and show that the schedule of presenting data during training is very important. Second, we develop a stacked architecture that includes warping of the second image with intermediate optical flow. Third, we elaborate on small displacements by introducing a sub-network specializing on small motions. FlowNet 2.0 is only marginally slower than the original FlowNet but decreases the estimation error by more than 50%. It performs on par with state-of-the-art methods, while running at interactive frame rates. Moreover, we present faster variants that allow optical flow computation at up to 140fps with accuracy matching the original FlowNet.
Publication titleFlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks
Publication authorsEddy Ilg, Nikolaus Mayer, Tonmoy Saikia, Margret Keuper, Alexey Dosovitskiy, Thomas Brox
Publication venueCVPR 2017
Publication URLhttps://arxiv.org/pdf/1612.01925
Input Data TypesUses Depth Input
Programming language(s)Python
HardwareGeForce 1080Ti
Websitehttps://lmb.informatik.uni-freiburg.de/Publications/2017/IMSKDB17/
Source code or download URLhttps://github.com/NVIDIA/flownet2-pytorch
Submission creation date17 Mar, 2020
Last edited17 Mar, 2020

Optical Flow

Method InfoAccuracy (<20px)EPE (pixel)
FlowNet 2.0copyleft0.68527.610