X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=blobdiff_plain;f=mtp_example.cc;fp=mtp_example.cc;h=13cd4f106d0e1d4a4ccca22dbdd2d8232d0ebb8d;hb=9665f6b8d842f02facda1590200742482d4362fa;hp=0000000000000000000000000000000000000000;hpb=4d9b58034ce82094c233b61da247e11a584ec0bd;p=mtp.git

diff --git a/mtp_example.cc b/mtp_example.cc
new file mode 100644
index 0000000..13cd4f1
--- /dev/null
+++ b/mtp_example.cc
@@ -0,0 +1,156 @@
+
+///////////////////////////////////////////////////////////////////////////
+// This program is free software: you can redistribute it and/or modify  //
+// it under the terms of the version 3 of the GNU General Public License //
+// as published by the Free Software Foundation.                         //
+//                                                                       //
+// This program is distributed in the hope that it will be useful, but   //
+// WITHOUT ANY WARRANTY; without even the implied warranty of            //
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU      //
+// General Public License for more details.                              //
+//                                                                       //
+// You should have received a copy of the GNU General Public License     //
+// along with this program. If not, see <http://www.gnu.org/licenses/>.  //
+//                                                                       //
+// Written by and Copyright (C) Francois Fleuret                         //
+// Contact <francois.fleuret@idiap.ch> for comments & bug reports        //
+///////////////////////////////////////////////////////////////////////////
+
+// Multi-Tracked Path
+
+#include <iostream>
+#include <fstream>
+
+using namespace std;
+
+#include "tracker.h"
+
+//////////////////////////////////////////////////////////////////////
+
+scalar_t detection_score(scalar_t a, scalar_t b, scalar_t score_noise, scalar_t flip_noise) {
+  if(drand48() > flip_noise) {
+    return a + score_noise * (2.0 * drand48() - 1.0);
+  } else {
+    return b + score_noise * (2.0 * drand48() - 1.0);
+  }
+}
+
+int main(int argc, char **argv) {
+  int nb_locations = 7;
+  int nb_time_steps = 8;
+  int motion_amplitude = 1;
+
+  Tracker *tracker = new Tracker();
+
+  tracker->allocate(nb_time_steps, nb_locations);
+
+  // We define the spatial structures by stating what are the possible
+  // motions of targets, and what are the entrances and the
+  // exits.
+
+  // Here our example is a 1D space with motions from any location to
+  // any location less than motion_amplitude away, entrance at
+  // location 0 and exit at location nb_locations-1.
+
+  for(int l = 0; l < nb_locations; l++) {
+    for(int k = 0; k < nb_locations; k++) {
+      tracker->allowed_motion[l][k] = abs(l - k) <= motion_amplitude;
+    }
+    tracker->entrances[0] = 1;
+    tracker->exits[nb_locations - 1] = 1;
+  }
+
+  // We construct the graph corresponding to this structure
+
+  tracker->build_graph();
+
+  // Then, we specify for every location and time step what is the
+  // detection score there.
+
+  scalar_t flip_noise = 0.05;
+  scalar_t score_noise = 0.0;
+
+  // We first put a background noise, with negative scores at every
+  // location.
+
+  for(int t = 0; t < nb_time_steps; t++) {
+    for(int l = 0; l < nb_locations; l++) {
+      tracker->detection_scores[t][l] = detection_score(-1.0, 1.0, score_noise, flip_noise);
+    }
+  }
+
+  // Then we two targets with the typical local minimum:
+  //
+  // * Target A moves from location 0 to the middle, stays there for a
+  //   while, and comes back, and is strongly detected on the first
+  //   half
+  //
+  // * Target B moves from location nb_locations-1 to the middle, stay
+  //   there for a while, and comes back, and is strongly detected on
+  //   the second half
+
+  int la, lb; // Target locations
+  scalar_t sa, sb; // Target detection scores
+  for(int t = 0; t < nb_time_steps; t++) {
+    if(t < nb_time_steps/2) {
+      la = t;
+      lb = nb_locations - 1 - t;
+      sa = detection_score(10.0, -1.0, score_noise, flip_noise);
+      sb = detection_score( 1.0, -1.0, score_noise, flip_noise);
+    } else {
+      la = nb_time_steps - 1 - t;
+      lb = t - nb_time_steps + nb_locations;
+      sa = detection_score( 1.0, -1.0, score_noise, flip_noise);
+      sb = detection_score(10.0, -1.0, score_noise, flip_noise);
+    }
+
+    if(la > nb_locations/2 - 1) la = nb_locations/2 - 1;
+    if(lb < nb_locations/2 + 1) lb = nb_locations/2 + 1;
+
+    tracker->detection_scores[t][la] = sa;
+    tracker->detection_scores[t][lb] = sb;
+  }
+
+  // Does the tracking per se
+
+  {
+    ofstream out_tracker("/tmp/tracker.dat");
+    tracker->write(&out_tracker);
+
+    ifstream in_tracker("/tmp/tracker.dat");
+    Tracker tracker2;
+    tracker2.read(&in_tracker);
+    tracker2.build_graph();
+    tracker2.track();
+    ofstream out_traj("/tmp/result.trj");
+    tracker2.write_trajectories(&out_traj);
+  }
+
+  tracker->track();
+
+  // Prints the detected trajectories
+
+  for(int t = 0; t < tracker->nb_trajectories(); t++) {
+    cout << "TRAJECTORY "
+         << t
+         << " [starting " << tracker->trajectory_entrance_time(t)
+         << ", score " << tracker->trajectory_score(t) << "]";
+    for(int u = 0; u < tracker->trajectory_duration(t); u++) {
+      cout << " " << tracker->trajectory_location(t, u);
+    }
+    cout << endl;
+  }
+
+  // Save the underlying graph in the dot format, with occupied edges
+  // marked in bold.
+
+  {
+    ofstream dot("graph.dot");
+    tracker->print_graph_dot(&dot);
+    cout << "Wrote graph.dot." << endl;
+  }
+
+  delete tracker;
+
+  exit(EXIT_SUCCESS);
+}