RAFT: Recurrent All-Pairs Field Transforms for Optical Flow: Difference between revisions

(Created page with "RAFT: Recurrent All-Pairs Field Transforms for Optical Flow (ECCV 2020) * [https://arxiv.org/abs/2003.12039 Arxiv mirror] * [https://github.com/princeton-vl/RAFT Github repo]...")
 
 
Line 58: Line 58:
Note that the inputs are not shown clearly in the main figure. Th flow is also an input to the network.   
Note that the inputs are not shown clearly in the main figure. Th flow is also an input to the network.   
In their code the inputs are:
In their code the inputs are:
* <code>net</code> = the hidden features
* <code>net</code> the hidden features
* <code>inp</code> = the context
* <code>inp</code> the context
* <code>corr</code> = the correlation
* <code>corr</code> the correlation
* <code>flow</code> = the flow.
* <code>flow</code> the flow.
The outputs are:
* <code>net</code> the new hidden features
* <code>mask</code> the mask for upsampling the optical flow
* <code>delta_flow</code> the flow update
 
Their update block consists of two ConvGRU units.
See their paper and code for details.


==Evaluation==
==Evaluation==