This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
Answer my own question... So yeah I am right, tracking and detection are two different topics. Detection, in most cases, are in general high CPU-consuming where trackers are considered less CPU consuming i.e.simply centroid tracker, or other implementation you can find in opencv_contrib/tracking like kcf, moose, or much more advanced that uses neural network GO TURN, are for tracking problem, you can detect an object with object detector like Haar cascades, HOG + Linear SVM, or just thresholding, counter it doesn't make difference. The main thing is that you init tracker with ROI or box(points) and for the updated frame, it returns new bounding box, that should contain your tracked object. A tracker can get make mistakes over time, so it is good practice to seed it again with new ROI of an object provided by object detector algorithms, or any other way like even simply selection ROI on an image. I hope that helps and not only me had that problem. ;)
| | 2 | No.2 Revision |
Answer my own question... So yeah I am right, tracking Tracking and detection are two different topics. Detection, in most cases, are in general high CPU-consuming where trackers are considered less CPU consuming i.e.simply centroid tracker, or other implementation you can find in opencv_contrib/tracking like kcf, moose, or much more advanced that uses neural network GO TURN, are for tracking problem, you can detect an object with object detector like Haar cascades, HOG + Linear SVM, or just thresholding, counter it doesn't make difference. The main thing is that you init tracker with ROI or box(points) and for the updated frame, it returns new bounding box, that should contain your tracked object. A tracker can get make mistakes over time, so it is good practice to seed it again with new ROI of an object provided by object detector algorithms, or any other way like even simply selection ROI on an image. I hope that helps and not only me had that problem. ;)
| | 3 | No.3 Revision |
Answer my own question... Tracking and detection are two different topics. Detection, in most cases, are in general high CPU-consuming where trackers are considered less CPU consuming i.e.simply centroid tracker, or other implementation you can find in opencv_contrib/tracking like kcf, moose, or much more advanced that uses neural network GO TURN, are for tracking problem, you can detect an object with object detector like Haar cascades, HOG + Linear SVM, or just thresholding, counter contour it doesn't make difference. The main thing is that you init tracker with ROI or box(points) and for the updated frame, it returns new bounding box, that should contain your tracked object. A tracker can get make mistakes over time, so it is good practice to seed it again with new ROI of an object provided by object detector algorithms, or any other way like even simply selection ROI on an image. I hope that helps and not only me had that problem. ;)
