You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello, I am very interested in your work, but I have a question to ask you. In the RPN stage, a proposal may be a positive sample in dataset A but a negative sample in dataset B. Will this produce ambiguity? Thanks!
The text was updated successfully, but these errors were encountered:
Great question! You are correct. We did not handle this issue in our work. We tried a few ideas including dataset-specific RPN, class-specific RPN, and ignoring high confident losses in RPN. All of them did not improve or improved minor.
My hypothesis is that RPNs are robust enough to generalize even if they are not trained consistently, given that we used an over-sufficient number of proposals at testing (1K proposals per image). You can find more discussions about RPN generalization in Appendix A of our Detic project.
Hello, I am very interested in your work, but I have a question to ask you. In the RPN stage, a proposal may be a positive sample in dataset A but a negative sample in dataset B. Will this produce ambiguity? Thanks!
The text was updated successfully, but these errors were encountered: