The method should work okay but will probably require a new config file and/or an object model. In particular, I think the object model could be very helpful.
One note for future improvements: bayes_extract_tracker_data() is easily the bottle neck in this function. It uses np.isin(), which I'm pretty sure runs in linear time, O(n). There has to be a faster way to accomplish what I want, possibly
by using numba.njit. See here for a somewhat related answer. Though I need the indices of the located values, not just the values themselves, so those answers might not be applicable.
This PR is mainly to include a new tracking method from: https://github.com/quantumjot/BayesianTracker
The method should work okay but will probably require a new config file and/or an object model. In particular, I think the object model could be very helpful.
One note for future improvements:
bayes_extract_tracker_data()is easily the bottle neck in this function. It usesnp.isin(), which I'm pretty sure runs in linear time,O(n). There has to be a faster way to accomplish what I want, possibly by usingnumba.njit. See here for a somewhat related answer. Though I need the indices of the located values, not just the values themselves, so those answers might not be applicable.Minor change:
verboseoption toPipelineandOrchestrator