knazeri / edge-connect

EdgeConnect: Structure Guided Image Inpainting using Edge Prediction, ICCV 2019 https://arxiv.org/abs/1901.00212
http://openaccess.thecvf.com/content_ICCVW_2019/html/AIM/Nazeri_EdgeConnect_Structure_Guided_Image_Inpainting_using_Edge_Prediction_ICCVW_2019_paper.html
Other
2.5k stars 530 forks source link

compared with CA version2:deepFill v2 #113

Open tangeroo opened 5 years ago

tangeroo commented 5 years ago

Hi, Sorry to bother.But I'm really curious about which is better when compared with deepFill v2.Have you ever tested that?

neouyghur commented 5 years ago

Hi,

I don't know which one is better, but edge-connect network design should be efficient than deepfill v2, and I think using gated convolution is not so critical since dilated convolution used in most of the layers.

On Fri, Sep 6, 2019 at 7:13 PM tangeroo notifications@github.com wrote:

Hi, Sorry to bother.But I'm really curious about which is better when compared with deepFill v2.Have you ever tested that?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/knazeri/edge-connect/issues/113?email_source=notifications&email_token=ABQRQ3BQEFKW3WCMRDVG6LLQIINLNA5CNFSM4IUGZOF2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HJYERAQ, or mute the thread https://github.com/notifications/unsubscribe-auth/ABQRQ3FO2F7EXQQE4NXYTL3QIINLNANCNFSM4IUGZOFQ .