mAP
mAP copied to clipboard
Cartucho
I want to display the bbox of FN, what should I do
I wanted to look at the results of FN, and I saw in the code that it was prompted to use the orange box to mark FN, but I didn't see the results when I run the code.
Just like this picture, the bbox of TP is shown.But the FN in the upper right corner is not shown. How can I modify the code? Thank you
Hello! That is a great idea, I will add it to my TODO list. I am currently restructuring the full code and the False Negatives are not being plotted yet.
I will add it as soon as possible!
I just added it! If you update it, it should now draw the false negatives in pink.
Man!!!! Thank tou Very much! I'm working in a object detection project to use in roadways! My objective is detect animals! So show what objects present in gt wasn't detected is more important! I was very happy with the news! I'm citting your work! Thank you very much!
Em qui, 12 de mar de 2020 20:15, Cartucho [email protected] escreveu:
I just added it! If you update it, it should now draw the false negatives in pink.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Cartucho/mAP/issues/65#issuecomment-598473115, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHPAFKEJ5MCGLCS6KCPAV2TRHFUKJANCNFSM4G7ORQNQ .
@Cartucho Something is happening during draw FN bb. In my case, when some image has not any detection, the file doesn't have coordinates content. So that image is not saved in output. But is a false negative...so, the mAP is computing these situations? I believe that the algorithm is comparing detection bounding-boxes with ground-truth bounding boxes but if it isn't detection in an image, this file is not considered! For example, I have a 024.jpg file that has a ground-truth 024.txt, but the object wasn't detected and the 024.txt detection-result file is empty. When I run the evaluation, receive the bellow error, cause I believe that the image was not generated:
Traceback (most recent call last):
File "main.py", line 750, in
@Cartucho here we got a false negative with the pink bounding box but at the output result, there are false-positive and true positive numbers, how to count false negative?
