Not getting results from cv2.dnn.readNetFromDarknet() [closed]

asked 2018-04-24 07:43:21 -0600

I m working on darknet with openCV in python. when I try to run the code with camera or video I am getting all the output values to zero as follows: (0.34697276, 0.04259526, 1.2586619, 1.0548509, 0.0053665913, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0)

Here, All the class confidence are zeros.

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by dkurt
close date 2018-05-01 14:59:50.850149

add a comment

1 answer

Sort by » oldest newest most voted

answered 2018-04-24 10:33:25 -0600

dkurt

1424 ●7 ●17

@DeamonDevs, have you applied image pre-processing such mean subtraction or normalization? We don't even know what kind of model did you use but I guess you're trying one of YOLOs. Note that all the YOLO networks work with images with values normalized to [0, 1] range (source: https://github.com/opencv/opencv/tree...).