Interponet, a brain inspired neural network for optical flow dense interpolation
S Zweig, L Wolf - Proceedings of the IEEE Conference on …, 2017 - openaccess.thecvf.com
Sparse-to-dense interpolation for optical flow is a fundamental phase in the pipeline of most
of the leading optical flow estimation algorithms. The current state-of-the-art method for
interpolation, EpicFlow, is a local average method based on an edge aware geodesic
distance. We propose a new data-driven sparse-to-dense interpolation algorithm based on a
fully convolutional network. We draw inspiration from the filling-in process in the visual
cortex and introduce lateral dependencies between neurons and multi-layer supervision into …
of the leading optical flow estimation algorithms. The current state-of-the-art method for
interpolation, EpicFlow, is a local average method based on an edge aware geodesic
distance. We propose a new data-driven sparse-to-dense interpolation algorithm based on a
fully convolutional network. We draw inspiration from the filling-in process in the visual
cortex and introduce lateral dependencies between neurons and multi-layer supervision into …
[PDF][PDF] InterpoNet, A brain inspired neural network for optical flow dense interpolation Supplementary
S Zweig, L Wolf - openaccess.thecvf.com
Method KITTI 2012 KITTI 2015 EPE% Out-all EPE% Out-all FF+ Ours 2.363 11.11 7.921
29.00 FF+ Epic 3.518 11.25 16.100 33.00 CPM+ Ours 2.271 11.3 6.92 26.04 CPM+ Epic
3.337 11.16 15.135 32.48 DF+ Ours 2.074 9.01 6.626 24.29 DF+ Epic 2.92 12.34 11.680
30.34 DM+ Ours 2.168 9.57 6.733 28.84 DM+ Epic 3.515 14.20 14.068 35.12
29.00 FF+ Epic 3.518 11.25 16.100 33.00 CPM+ Ours 2.271 11.3 6.92 26.04 CPM+ Epic
3.337 11.16 15.135 32.48 DF+ Ours 2.074 9.01 6.626 24.29 DF+ Epic 2.92 12.34 11.680
30.34 DM+ Ours 2.168 9.57 6.733 28.84 DM+ Epic 3.515 14.20 14.068 35.12
Beste Ergebnisse für diese Suche Alle Ergebnisse