imglab: chinese ("automatic") clustering, keyboard shortcuts for zooming (#2007)

* imglab: add support for using chinese whispers for more automatic clustering

* widgets: refactor out zooming from wheel handling

* tools/imglab/src/metadata_editor.cpp

imglab: add keyboard shortcuts for zooming

dlib/gui_widgets/widgets.cpp

@@ -7065,25 +7065,9 @@ namespace dlib
 // ----------------------------------------------------------------------------------------
 
     void image_display::
-    on_wheel_up (
-        unsigned long state
+    zoom_in (
     )
     {
-        // disable mouse wheel if the user is drawing a rectangle
-        if (drawing_rect)
-            return;
-
-        // if CONTROL is not being held down
-        if ((state & base_window::CONTROL) == 0)
-        {
-            scrollable_region::on_wheel_up(state);
-            return;
-        }
-
-        if (rect.contains(lastx,lasty) == false || hidden || !enabled)
-            return;
-
-
         if (zoom_in_scale < 100 && zoom_out_scale == 1)
         {
             const point mouse_loc(lastx, lasty);
@@ -7119,7 +7103,7 @@ namespace dlib
 // ----------------------------------------------------------------------------------------
 
     void image_display::
-    on_wheel_down (
+    on_wheel_up (
         unsigned long state
     )
     {
@@ -7130,14 +7114,22 @@ namespace dlib
         // if CONTROL is not being held down
         if ((state & base_window::CONTROL) == 0)
         {
-            scrollable_region::on_wheel_down(state);
+            scrollable_region::on_wheel_up(state);
             return;
         }
 
         if (rect.contains(lastx,lasty) == false || hidden || !enabled)
             return;
 
+        zoom_in();
+    }
 
+// ----------------------------------------------------------------------------------------
+
+    void image_display::
+    zoom_out (
+    )
+    {
         if (zoom_in_scale != 1)
         {
             const point mouse_loc(lastx, lasty);
@@ -7170,6 +7162,30 @@ namespace dlib
         }
     }
 
+// ----------------------------------------------------------------------------------------
+
+    void image_display::
+    on_wheel_down (
+        unsigned long state
+    )
+    {
+        // disable mouse wheel if the user is drawing a rectangle
+        if (drawing_rect)
+            return;
+
+        // if CONTROL is not being held down
+        if ((state & base_window::CONTROL) == 0)
+        {
+            scrollable_region::on_wheel_down(state);
+            return;
+        }
+
+        if (rect.contains(lastx,lasty) == false || hidden || !enabled)
+            return;
+
+        zoom_out();
+    }
+
 // ----------------------------------------------------------------------------------------
 // ----------------------------------------------------------------------------------------
 //         image_window member functions

dlib/gui_widgets/widgets.h

@@ -3493,6 +3493,12 @@ namespace dlib
         bool overlay_editing_is_enabled (
         ) const { auto_mutex M(m); return overlay_editing_enabled; }
 
+        void zoom_in (
+        );
+
+        void zoom_out (
+        );
+
     private:
 
         void draw (

tools/imglab/src/cluster.cpp

@@ -10,6 +10,7 @@
 #include <dlib/dir_nav.h>
 #include <dlib/clustering.h>
 #include <dlib/svm.h>
+#include <dlib/statistics.h>
 
 // ----------------------------------------------------------------------------------------
 
@@ -72,6 +73,56 @@ std::vector<assignment> angular_cluster (
     }
     return assignments;
 }
+std::vector<assignment> chinese_cluster (
+    std::vector<matrix<double,0,1> > feats,
+    unsigned long &num_clusters
+    )
+{
+    // try to find a good value to select if we should add a vertex in the graph
+    matrix<double,0,1> m;
+    for (unsigned long i = 0; i < feats.size(); ++i)
+        m += feats[i];
+    m /= feats.size();
+
+    for (unsigned long i = 0; i < feats.size(); ++i)
+    {
+        feats[i] -= m;
+        double len = length(feats[i]);
+        if (len != 0)
+            feats[i] /= len;
+    }
+
+    running_stats<double> rs;
+    for (size_t i = 0; i < feats.size(); ++i) {
+        for (size_t j = i; j < feats.size(); ++j) {
+            rs.add(length(feats[i] - feats[j]));
+        }
+    }
+
+    // add vertices for chinese whispers to find clusters
+    std::vector<sample_pair> edges;
+    for (size_t i = 0; i < feats.size(); ++i) {
+        for (size_t j = i; j < feats.size(); ++j) {
+            if (length(feats[i] - feats[j]) < rs.mean()) {
+                edges.push_back(sample_pair(i, j, length(feats[i] - feats[j])));
+            }
+        }
+    }
+
+    std::vector<unsigned long> labels;
+    num_clusters = chinese_whispers(edges, labels);
+
+    std::vector<assignment> assignments;
+    for (unsigned long i = 0; i < feats.size(); ++i)
+    {
+        assignment temp;
+        temp.c = labels[i];
+        temp.dist = length(feats[i]);
+        temp.idx = i;
+        assignments.push_back(temp);
+    }
+    return assignments;
+}
 
 // ----------------------------------------------------------------------------------------
 
@@ -134,7 +185,7 @@ int cluster_dataset(
         return EXIT_FAILURE;
     }
 
-    const unsigned long num_clusters = get_option(parser, "cluster", 2);
+    unsigned long num_clusters = get_option(parser, "cluster", 0);
     const unsigned long chip_size = get_option(parser, "size", 8000);
 
     image_dataset_metadata::dataset data;
@@ -177,7 +228,12 @@ int cluster_dataset(
     }
 
     cout << "\nClustering objects..." << endl;
-    std::vector<assignment> assignments = angular_cluster(feats, num_clusters);
+    std::vector<assignment> assignments;
+    if (num_clusters) {
+        assignments = angular_cluster(feats, num_clusters);
+    } else {
+        assignments = chinese_cluster(feats, num_clusters);
+    }
 
 
     // Now output each cluster to disk as an XML file.

tools/imglab/src/main.cpp

@@ -588,7 +588,7 @@ int main(int argc, char** argv)
                                         "The parts are instead simply mirrored to the flipped dataset.", 1);
         parser.add_option("rotate", "Read an XML image dataset and output a copy that is rotated counter clockwise by <arg> degrees. "
                                   "The output is saved to an XML file prefixed with rotated_<arg>.",1);
-        parser.add_option("cluster", "Cluster all the objects in an XML file into <arg> different clusters and save "
+        parser.add_option("cluster", "Cluster all the objects in an XML file into <arg> different clusters (pass 0 to find automatically) and save "
                                      "the results as cluster_###.xml and cluster_###.jpg files.",1);
         parser.add_option("ignore", "Mark boxes labeled as <arg> as ignored.  The resulting XML file is output as a separate file and the original is not modified.",1);
         parser.add_option("rmlabel","Remove all boxes labeled <arg> and save the results to a new XML file.",1);
@@ -704,7 +704,7 @@ int main(int argc, char** argv)
         parser.check_incompatible_options("box-images", "ignore");
         const char* convert_args[] = {"pascal-xml","pascal-v1","idl"};
         parser.check_option_arg_range("convert", convert_args);
-        parser.check_option_arg_range("cluster", 2, 999);
+        parser.check_option_arg_range("cluster", 0, 999);
         parser.check_option_arg_range("rotate", -360, 360);
         parser.check_option_arg_range("size", 10*10, 1000*1000);
         parser.check_option_arg_range("min-object-size", 1, 10000*10000);

tools/imglab/src/metadata_editor.cpp

@@ -343,6 +343,16 @@ on_keydown (
             last_keyboard_jump_pos_update = 0;
         }
 
+        if (key == '=')
+        {
+            display.zoom_in();
+        }
+
+        if (key == '-')
+        {
+            display.zoom_out();
+        }
+
         if (key == 'd' && (state&base_window::KBD_MOD_ALT))
         {
             remove_selected_images();