search

yzcjtr / GeoNet

Code for GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose (CVPR 2018)
MIT License
723 stars 181 forks source link

figure 5 in paper #11

Closed tongpinmo closed 6 years ago

tongpinmo commented 6 years ago

hello ,in Figure 5 of your paper,GeoNet shows clear advantages in occluded ,texture ambiguous regions,and even in shaded dim area. but i can't see the comparison of GeoNet Predictions and other methods like DirFlowNetsS,can you show it in the figure ,and where is the difference between GeoNet Error and DirFlowNetsS Error? i can't see the advantages of GeoNet.can you show the accurate data and show more pictures (not just chose the best 4 input)

yzcjtr commented 6 years ago

In Fig.5, the 4th & 5th columns show the qualitative comparison between GeoNet and DirFlowNetS. The red or orange regions indicate erroneous predictions. Can't you see the difference of error distributions? Regarding accurate quantitative comparison, refer to Table.2 where we show the end point error statistics, and in Fig.6 we analysed the EPE distribution over the norm of groundtruth residual flow.

tongpinmo commented 6 years ago

oh,oh ,sorry ,i don' t make clear before ,thank you for your explanation.it means the bigger red and orange regions ,the more erroneous predictions are ?

yzcjtr commented 6 years ago

Yes. Such visualization approach follows the official KITTI benchmark, as provided in their development kit.