-
Notifications
You must be signed in to change notification settings - Fork 141
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
about DetEval.py evaluation speed boosting. #16
Comments
Oh yes you are most welcome, thanks for wanting to contribute! And I assume that you have already tested out your code? In particular, does the library provides an exact solution instead of approximation when it comes to calculating the area of intersection? |
I have compared the result of mask counting method and polygon based method, their result have a bit difference(1 to 3 pixels in computing iou between 2 polygons with area about 10000 pixels) because mask counting based implement discretize the coordinates to pixels. In fact, I think the polygon based method produce more precise result. |
Yup that makes sense. Also, less than 1 percent of differences is probably negligible in practice. |
Thx for your replying, I will make a pull request soon! 😆 |
Hi, I found that the evaluation in this code run extremely slowly and most time-consuming operation in your code is area/area_of_intersection/iou. These functions are based on mask counting, which depends highly on the size of images(some big-size images can be bottleck of computing).
I have replaced mask counting operation with polygon coordinate computing(which uses shapely, a geometry lib written in python) so that it highly boosts the evaluatoin process.
Can I make a PR? Hope for your replying, thx.
The text was updated successfully, but these errors were encountered: