Ask Your Question

Cascade classier DetectMultiScale()

asked 2018-03-09 07:47:31 -0600

Jad gravatar image

So am trying to use cascade.detectMultiScale() as a classifier by using an input image of size 24*24 (the same size used in the cascade training) and the parameters:

detections= my_cascade.detectMultiScale(img, scaleFactor=1.05, minNeighbors=1, minSize=(1,1), maxSize=(24,24))

I am not getting any detections at all although I am using the same positive images used in the training to test! What is causing that!

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2018-03-09 08:28:35 -0600

berak gravatar image

updated 2018-03-09 08:55:21 -0600

24x24 is the smallest rect, that can be detected (the training window of the cascade)

so, minSize=(1,1), is useless (to small) , and maxSize=(24,24)) will forbid to detect it.

also, you probably need some border around it in the detection image

so maybe try this detections= my_cascade.detectMultiScale(img, scaleFactor=1.05, minNeighbors=1) first and then try to finetune.

edit flag offensive delete link more

Question Tools

1 follower


Asked: 2018-03-09 07:47:31 -0600

Seen: 367 times

Last updated: Mar 09 '18