Hi, I am using CV2 for Custom Model Detections. Can anyone tell how to use multiple custom models in one Program. Like in Fruit Detection, if I made custom cascades of Banana, Apple, and Oranges. Now how I gonna use these three cascades in one Program. Thanks
you instantiate three classifiers, each loading one cascade?
related:
Yes, I instantiate three classifiers, each loading one cascade, and also this is me on StackOverflow:
terrible idea. do not try to abuse a detection algo for classification. instead try with a cnn like SSD or yolo
Is it possible with OpenCV?
My idea is storing value of each cascade in lists and others values like detection, confidence, roi and others in sub lists. And finally display one with higher condifence.
unfortunately, the “confidence” values of different cascades are not comparable in a meaningful way.
again, rather have a look at yolo which should do all you wanted, and is nicely supported from opencv’s dnn module
Shifting to Yolo. In YOLO, Single Custom Object Detection Seems Quite Simple. What about Multiple? Can we train multiple objects in one YOLO Weight file?
sure you can ! that’s the whole point …