0
$\begingroup$

The label output file of the yolo v5 model is something like this:

0 0.0917969 0.907639 0.172656 0.170833 0.475255
1 0.387109 0.15 0.127344 0.291667 0.822308
0 0.115625 0.539583 0.195312 0.354167 0.828012
0 0.906641 0.473611 0.177344 0.35 0.829128
0 0.198047 0.283333 0.167969 0.322222 0.846307
0 0.796484 0.294444 0.160156 0.316667 0.866312
0 0.642187 0.179167 0.123438 0.272222 0.88274
  • This text file is one of the results for the labels from detect.py. I know for sure that the 1st column is related to the class. As there are no headings in the text file, how can I know which these other columns are relating to. Values being normalized, it is difficult to understand which columns are for xmin, ymin, xmax, ymax, confidence respectively.
  • Is there any method that can be used in the detect.py to exclude the confidence column from this annotations text file.
  • Also, I am little confused here because yolov5 accepts (xcentre, y centre, width, height) as input values of bounding boxes but, I guess, this output is in the form of (xmin, ymin, xmax, ymax). Does yolov5 automatically makes this change.
$\endgroup$
1
  • $\begingroup$ the output is cx cy width height score. $\endgroup$
    – sydnash
    Nov 24 at 3:03

1 Answer 1

0
$\begingroup$

--save-conf argument in detect.py command is the solution.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.