Closed xxxzhi closed 4 years ago
Sorry for the typo! By the way, the misleading result comes from our another work: PaStaNet on CVPR2020, where we perform late fusion on iCAN and PaStaNet thus improving the performance of iCAN. Sorry again for the typo and thanks for pointing out!
Thanks for your reply. Btw, I suppose you also have the result of HAKE-3D with PaStaNet. Would mind provide the result?
The combination of HAKE-3D and PaStaNet is still in progress.
Hi, in the readme file, you provide the iCAN result as follows,
However, in the paper, the result of iCAN is,
Would u mind provide the details of your new iCAN result?