Marking YOLOv5 Detection Text Outputs with TP or FP

[kyrangraves]

Hi All,

I have trained a YOLOv5 model and used val.py to test the performance of the trained model using independent labelled datasets. I have set the val.py script to write detections to text files. Is there a way to extract/denote for each detection made whether or not the detection is a true positive or false positive? This information is surely determined for each detection to calculate performance metrics. Is there a way to easily output this information in the text files?

I would ideally like to produce a text-output like this,

5 0.794286 0.599206 0.0373207 0.0625995 0.956099 TP
3 0.652693 0.639765 0.0693291 0.138997 0.847412 TP
3 0.954316 0.848078 0.0913681 0.159411 0.170322 FP

I have looked through the discussion threads and elsewhere online and I seemingly cannot find any answers. This may be because it's obvious, or just not yet been discussed. Any information would be greatly appreciated!

Thank you,
Kyran