How to export the prediction(boundingbox x,y,w,h, class, e.g) in a txt file?

[dsx0511]

Hallo, I have already trained a model for my custom object and it works well. After the Detection I can get a picture with the predicted bounding box named "prediction". Now I have to evaluate this trainded model using some metics by comparison with the ground truth. So what I need are also the exact value of x, y, w, h. My Question is: How can I get these values in a txt-file by modifying some code or using some command? Thanks a lot!

[ahsan856jalal]

https://github.com/pjreddie/darknet/pull/640

Regards Ahsan

On Wed, Apr 25, 2018 at 2:19 PM, Shuxiao Ding notifications@github.com wrote:

Hallo, I have already trained a model for my custom object and it works well. After the Detection I can get a picture with the predicted bounding box named "prediction". Now I have to evaluate this trainded model using some metics by comparison with the ground truth. So what I need are also the exact value of x, y, w, h. My Question is: How can I get these values in a txt-file by modifying some code or using some command? Thanks a lot!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/AlexeyAB/darknet/issues/694, or mute the thread https://github.com/notifications/unsubscribe-auth/AK9zjqf53lNFS4MC_8lVpn7IT-QvTXVfks5tsD-EgaJpZM4TjEC8 .

[AlexeyAB]

@dsx0511

Un-comment these 2 lines:

1. https://github.com/AlexeyAB/darknet/blob/15139eb6abd3c6c05a4046b5ac917de42c678827/src/image.c#L390

2. https://github.com/AlexeyAB/darknet/blob/15139eb6abd3c6c05a4046b5ac917de42c678827/src/image.c#L246

Re-compile. And run this command:

• For list of images in the data/train.txt file: darknet.exe detector test data/voc.data yolo-voc.cfg yolo-voc.weights -dont_show < data/train.txt > result.txt

• For videofile test.mp4: darknet.exe detector demo data/voc.data yolo-voc.cfg yolo-voc.weights test.mp4 > result.txt

The result will be in the file result.txt

[dsx0511]

Thank you for these solutions!

[lvshuaigg]

Excuse me, can you save the predicted picture? For example, I want to predict the picture in the valid, and save the picture after the prediction.