search

jiaw-z / FCStereo

[CVPR'22] Revisiting Domain Generalized Stereo Matching Networks from a Feature Consistency Perspective
MIT License
53 stars 3 forks source link

About the experiment results in your paper. #7

Closed Zhaohuai-L closed 1 year ago

Zhaohuai-L commented 1 year ago

Hi , Thanks a lot for your great work,Can I know how you tested in your Table1 about the PSMNet Results? image

It has different results from the following papers: Table3 in Domain-invariant Stereo Matching Networks image Table5 in GraftNet: Towards Domain Generalized Stereo Matching with a Broad-Spectrum and Task-Oriented Feature image

jiaw-z commented 1 year ago

Hi , Thanks a lot for your great work,Can I know how you tested in your Table1 about the PSMNet Results? image

It has different results from the following papers: Table3 in Domain-invariant Stereo Matching Networks image Table5 in GraftNet: Towards Domain Generalized Stereo Matching with a Broad-Spectrum and Task-Oriented Feature image

We implement the PSMNet and evaluate the domain generalization performance. We report results of 3-pixel error rate. Similar results are also observed in other works. So we decide to report the results of our implemented PSMNet. AcfNet: image STTR: image

Zhaohuai-L commented 1 year ago

OK, thanks for your reply.