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.

Sign up for GitHub

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

Jump to bottom

What metric do you use to test your own dataset? #101

Open
ba77-ku opened this issue Jan 15, 2022 · 4 comments
Open

What metric do you use to test your own dataset? #101

ba77-ku opened this issue Jan 15, 2022 · 4 comments

Comments

@ba77-ku
Copy link

ba77-ku commented Jan 15, 2022

We trained from scratch with our own dataset, but the DexiNED prediction results are not binary images like our labels. What method do you use to test tags with predictive images? Are you converting by skeletonizing or specifying a threshold? Because when I use the f-score metric I get very poor results. Additionally, what is the accuracy metric you are using?
Output and label images are below.
image_10000_average
image_10000
@xavysp
Copy link
Owner

xavysp commented Jan 23, 2022

Hi, welcome on board.
did you check issue: Evaluation index of edge detection #32

@xavysp
Copy link
Owner

xavysp commented Jan 23, 2022

Let me know if you have further questions

@xavysp xavysp mentioned this issue Jan 31, 2022
Closed
@ba77-ku
Copy link
Author

ba77-ku commented Feb 7, 2022

What is the number of epochs you would recommend during the training phase?

@xavysp
Copy link
Owner

xavysp commented Feb 24, 2022

Hi, it depends to the dataset, on BIPED for example you can stop in 9-13 epochs

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@xavysp @ba77-ku