This is a very simple implementation of the KSP applied to
-multi-target tracking. It is dubbed Multi-Tracked Path since it
+multi-target tracking. It is dubbed Multi-Tracked Paths since it
differs a bit from the standard KSP. It works with negative edge
-length and stops when it can not find path of negative length anymore
-instead of fixing the total number of paths to a constant.
+length and stops when it can not find any path of negative length,
+instead of fixing the total number of paths to a constant K.
The two main classes are MTPGraph and Tracker.
The MTPGraph class stores a directed acyclic graph (DAG), with a
-length for each edge (which can be negative), and can compute the
-family of paths in this graph that minimizes the sum of edge
-lengths.
+length for each edge -- which can be negative -- and can compute the
+family of paths in this graph that minimizes the sum of edge lengths.
This means that it will iteratively add paths as long as it can find
some with negative length. If there are no such path, it will compute
-no path at all. Note that the solution it finds is globally optimal.
+no path at all. Note that the solution it finds is globally
+optimal. Note that the procedure is similar to that of KSP, in the
+sense that the family it computes eventually is globally optimal, even
+if the procedure is iterative.
The Tracker class allows
(2) to define a number of time steps
- (3) to set for every location and time a detection score
+ (3) to set for every location and time a detection score, which
+ should be equal to log(P(Y = 1 | X)/P(Y = 0 | X)) where Y stands for
+ the location occupancy and X for the observations.
-From this input, it computes the best set of disjoint trajectories
+From this setting, it computes the best set of disjoint trajectories
consistent with the topology, which maximizes the overall detection
score (i.e. the sum of the detection scores of the nodes visited by
the trajectories)
projects / mtp.git / commitdiff