I'm recognizing objects on the Darknet using OpenCV. The image contains several objects of the same type. I need to assign an ID for each detected object, but cv::Tracker and cv::legacy::MultipleTracker require an image to work. I can't send object coordinates to them.
Is there a way in OpenCV to track (assign an ID and follow it in a sequence of frames) already detected objects?
I have cv::Rect for each object in image. But i didn't find way to send it to cv::Tracker or cv::MultiTracker.