Closed elliestath closed 2 years ago
Hi
Yes, Jaccard index and Intersection-Over-Union (IoU) are equal (in classification metrics) Please, feel free to let me know if you require any further information.
Thank you for the answer, very useful to clarify my doubt. I am performing a semantic segmentation task and would like to make sure I am using the correct metrics from your library. For now I do:
pixel accuracy: cm.ACC
precision: cm.PPV
recall: cm.TPR
F1-score: cm.F1
IoU: cm.J
does this sound right? Thank you very much, the library is awesome!
pixel accuracy: cm.ACC 🆗 (For overall accuracy take a look at cm.Overall_ACC ) precision: cm.PPV 🆗 recall: cm.TPR 🆗 F1-score: cm.F1 🆗 IoU: cm.J 🆗
Description
Hey, it works like a charm, thanks! Question: If I am not wrong, there is no IoU score but only the Jaccard index which seems to be equivalent. Can you confirm this?
Thanks!
PyCM Version (Use :
pycm.__version__
)pycm==3.4