Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Loss and missing detections #1036

Open
BigMuscle85 opened this issue Apr 19, 2023 · 0 comments
Open

Loss and missing detections #1036

BigMuscle85 opened this issue Apr 19, 2023 · 0 comments

Comments

@BigMuscle85
Copy link

I have a question if MultiBoxLoss function reflects false negatives (and possibly false positives)?

I'm trying to optimize my SSD detection model and while my optimizations provide lower training loss and higher testing accuracy, the result have more missing detections than before.

The multibox loss used in the SSD method is defined as L = Lconf + alpha * Lloc, but what happens when the labelled object is not detected at all? How is it reflected in the computed loss? Does it increase the loss or is it completely ignored?

I'm thinking to update the loss function to be something like L = Lconf + alpha * Lloc + beta * abs(number_of_detections - number_of_gt), but I don't know if it makes any sense.

Thank you.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant