About the algorithm model

[aha831]

Hello, I have been paying close attention to your new research, and thank you very much for the help your research has brought me. I have studied the nsff you proposed before and admire your new model DynIBaR for surpassing the nsff. I would like to ask you what is the core reason why DynIBaR can surpass nsff, is this ibrnet idea better than the nsff based on optical flow? If so, why does this ibrnet approach have this effect? Thank you very much for your answer