You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As previous algorithms did not give good results. So we searched for other algorithms.
Got one algorithm called "Tracking Learning Detection". Details for this algorithm can be found here
Understood "Tracking Learning Detection".
We searched for code of "Tracking Learning Detection" code and got one c++ code and windows exe here
Tested this exe on some objects and got problems like "If object changes its orientation it don't track object", "If object changes its side, it don't track the object".
As "Tracking Learning Detection" did not give fine results so we searched for other algorithms.
CMT Algorithm
Got one algorithm named "Consensus-based Matching and Tracking of Keypoints for Object Tracking (CMT)". Details for this algorithm can be found here
Understood working of CMT.
Gathered required knowledge for implementing CMT.
Implemented CMT in Python and OpenCV. Code can be found here
Tested CMT code using various objects and found "If object is moving it don't track object but if object stops, it gives bounding box", "If object escapes from the frame and comes back it don't track object again."