So there's this way of classifying an object as NotA (none of the above) class as described in https://arxiv.org/abs/1910.02830 So mostly this uses a sort of confidence thresholding with different approaches.

Can a classifier with such a NotA class be used as part of an object detection framework?

I'm assuming if the classifier isn't trained on a background class then NotA would "light up" also on background, so that class might need to be added as well to make some sense.

  • 1
    $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Sep 20, 2022 at 21:35
  • $\begingroup$ so a bot is asking me for clarifications? We live in the future, don't we $\endgroup$ Sep 21, 2022 at 13:28


You must log in to answer this question.

Browse other questions tagged .