From 1e1f4efbafd8571cd6c4849c7f4e9debdd0d000c Mon Sep 17 00:00:00 2001
From: Davis King <davis@dlib.net>
Date: Thu, 28 Mar 2013 19:21:52 -0400
Subject: [PATCH] Gave load_image_dataset() the ability to skip images that
 don't have any ground truth boxes.

---
 dlib/data_io/load_image_dataset.h          | 14 +++++++++----
 dlib/data_io/load_image_dataset_abstract.h | 24 +++++++++++++---------
 2 files changed, 24 insertions(+), 14 deletions(-)

diff --git a/dlib/data_io/load_image_dataset.h b/dlib/data_io/load_image_dataset.h
index fdb6945bb..c955cdefb 100644
--- a/dlib/data_io/load_image_dataset.h
+++ b/dlib/data_io/load_image_dataset.h
@@ -28,7 +28,8 @@ namespace dlib
         array<image_type,MM>& images,
         std::vector<std::vector<rectangle> >& object_locations,
         const std::string& filename,
-        const std::string& label 
+        const std::string& label,
+        bool skip_empty_images = false
     )
     {
         images.clear();
@@ -47,11 +48,10 @@ namespace dlib
         dataset data;
         load_image_dataset_metadata(data, filename);
 
-        images.resize(data.images.size());
+        image_type img;
         std::vector<rectangle> rects;
         for (unsigned long i = 0; i < data.images.size(); ++i)
         {
-            load_image(images[i], data.images[i].filename);
             rects.clear();
             for (unsigned long j = 0; j < data.images[i].boxes.size(); ++j)
             {
@@ -60,7 +60,13 @@ namespace dlib
                     rects.push_back(data.images[i].boxes[j].rect);
                 }
             }
-            object_locations.push_back(rects);
+
+            if (!skip_empty_images || rects.size() != 0)
+            {
+                object_locations.push_back(rects);
+                load_image(img, data.images[i].filename);
+                images.push_back(img);
+            }
         }
 
         set_current_dir(old_working_dir);
diff --git a/dlib/data_io/load_image_dataset_abstract.h b/dlib/data_io/load_image_dataset_abstract.h
index c31ff14c1..4800095b4 100644
--- a/dlib/data_io/load_image_dataset_abstract.h
+++ b/dlib/data_io/load_image_dataset_abstract.h
@@ -23,25 +23,29 @@ namespace dlib
         array<image_type,MM>& images,
         std::vector<std::vector<rectangle> >& object_locations,
         const std::string& filename,
-        const std::string& label 
+        const std::string& label,
+        bool skip_empty_images = false
     );
     /*!
         requires
             - image_type == is an implementation of array2d/array2d_kernel_abstract.h
             - pixel_traits<typename image_type::type> is defined  
         ensures
-            - This routine loads the images and their associated object boxes from 
-              the image metadata file indicated by filename.  This metadata file
-              should be in the XML format used by the save_image_dataset_metadata()
-              routine.
-            - #images.size() == the number of images in the metadata file
+            - This routine loads the images and their associated object boxes from the
+              image metadata file indicated by filename.  This metadata file should be in
+              the XML format used by the save_image_dataset_metadata() routine.
+            - #images.size() == The number of images loaded from the metadata file.  This
+              is all the images listed in the file unless skip_empty_images is set to true.
             - #images.size() == #object_locations.size()
-            - This routine is capable of loading any image format which can be read
-              by the load_image() routine.
+            - This routine is capable of loading any image format which can be read by the
+              load_image() routine.
             - for all valid i:  
-                - #images[i] == a copy of the ith image from the dataset
+                - #images[i] == a copy of the i-th image from the dataset
                 - #object_locations[i] == a vector of all the rectangles associated with
-                  #images[i].
+                  #images[i].  
+                - if (skip_empty_images == true) then
+                    - #object_locations[i].size() != 0
+                      (i.e. only images with detection boxes in them will be loaded.)
                 - if (labels != "") then
                     - only boxes with the given label will be loaded into object_locations.
                 - else