Hi AlexeyAB,
How to show only one label? Let's say for example,
I am using tiny-yolo model which detects 20 object classes. However instead of showing bounding box i.e. labels for all the object classes, is it possible to show only one object class? Let's say a picture contains Persons, dog, bicycle and cars. Once done it only shows bounding box for the person although it detects all the other objects.
Not sure if this will solve your issue but check it out https://github.com/AlexeyAB/darknet/issues/385#issuecomment-366102993
@rezaabdullah
Also the simplest way is to do these changes, if you want to detect persons only:
to this if (prob > thresh && class_id == 14) { if you use yolo-voc.weights (14 is number of line in the file voc.names where is person)
or to this if (prob > thresh && class_id == 0) { if you use yolo.weights (0 is number of line in the file coco.names where is person)
Tq
From: Alexey notifications@github.com
Sent: Thursday, March 8, 2018 5:23 AM
To: AlexeyAB/darknet
Cc: rezaabdullah; Mention
Subject: Re: [AlexeyAB/darknet] How to show one label only even though it detects all the objects (#433)
@rezaabdullahhttps://github.com/rezaabdullah
Also the simplest way is to do these changes, if you want to detect persons only:
And this line: https://github.com/AlexeyAB/darknet/blob/100d6f78011f0a773442411e2882a0203d390585/src/image.c#L251
to this if (prob > thresh && class_id == 14) { if you use yolo-voc.weights (14 is number of line in the file voc.names where is person)
or to this if (prob > thresh && class_id == 0) { if you use yolo.weights (0 is number of line in the file coco.names where is person)
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHubhttps://github.com/AlexeyAB/darknet/issues/433#issuecomment-371459838, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AffFLB0a_8uEu9frkU2ZMTQVJAQhoP6Bks5tcRTIgaJpZM4SiFzY.
Most helpful comment
@rezaabdullah
Also the simplest way is to do these changes, if you want to detect persons only:
to this
if (prob > thresh && class_id == 14) {if you use yolo-voc.weights (14 is number of line in the filevoc.nameswhere is person)or to this
if (prob > thresh && class_id == 0) {if you use yolo.weights (0 is number of line in the filecoco.nameswhere is person)