add FindCaffe cmake file for necessary dependency

modules/cnn_3dobj/CMakeLists.txt

@@ -1,3 +1,26 @@
-set(the_description "CNN for 3D object recognition and pose estimation including a completed Sphere View on 3D objects")
-ocv_define_module(cnn_3dobj opencv_core opencv_imgproc opencv_viz opencv_highgui caffe protobuf glog OPTIONAL WRAP python)
-target_link_libraries(opencv_cnn_3dobj caffe protobuf glog)
+set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} ${CMAKE_CURRENT_SOURCE_DIR})
+
+find_package(Caffe)
+if(Caffe_FOUND)
+  message(STATUS "Caffe:   YES")
+  set(HAVE_CAFFE 1)
+else()
+  message(STATUS "Caffe:   NO")
+endif()
+
+if(HAVE_CAFFE)
+configure_file(${CMAKE_CURRENT_SOURCE_DIR}/cnn_3dobj_config.hpp.in
+               ${CMAKE_CURRENT_SOURCE_DIR}/include/opencv2/cnn_3dobj_config.hpp @ONLY)
+
+include_directories(${CMAKE_CURRENT_BINARY_DIR})
+
+if(${Caffe_FOUND})
+  include_directories(${Caffe_INCLUDE_DIR})
+endif()
+  set(the_description "CNN for 3D object recognition and pose estimation including a completed Sphere View on 3D objects")
+  ocv_define_module(cnn_3dobj opencv_core opencv_imgproc opencv_viz opencv_highgui OPTIONAL WRAP python)
+
+if(${Caffe_FOUND})
+  target_link_libraries(opencv_cnn_3dobj ${Caffe_LIBS})
+endif()
+endif()

modules/cnn_3dobj/FindCaffe.cmake

@@ -0,0 +1,14 @@
+# Caffe package for CNN Triplet training
+unset(Caffe_FOUND)
+
+find_path(Caffe_INCLUDE_DIR NAMES caffe/caffe.hpp caffe/common.hpp caffe/net.hpp caffe/proto/caffe.pb.h caffe/util/io.hpp caffe/vision_layers.hpp
+  HINTS
+  /usr/local/include)
+
+find_library(Caffe_LIBS NAMES caffe
+  HINTS
+  /usr/local/lib)
+
+if(Caffe_LIBS AND Caffe_INCLUDE_DIR)
+    set(Caffe_FOUND 1)
+endif()

modules/cnn_3dobj/cnn_3dobj_config.hpp.in

@@ -0,0 +1,5 @@
+#ifndef __OPENCV_CNN_3DOBJ_CONFIG_HPP__
+#define __OPENCV_CNN_3DOBJ_CONFIG_HPP__
+// HAVE CAFFE
+#cmakedefine HAVE_CAFFE
+#endif

modules/cnn_3dobj/include/opencv2/cnn_3dobj.hpp

@@ -58,12 +58,16 @@ the use of this software, even if advised of the possibility of such damage.
 #include <tr1/memory>
 #include <dirent.h>
 #define CPU_ONLY
-#include "caffe/blob.hpp"
-#include "caffe/common.hpp"
-#include "caffe/net.hpp"
-#include "caffe/proto/caffe.pb.h"
-#include "caffe/util/io.hpp"
-#include "caffe/vision_layers.hpp"
+
+#ifdef HAVE_CAFFE
+#include <caffe/blob.hpp>
+#include <caffe/common.hpp>
+#include <caffe/net.hpp>
+#include <caffe/proto/caffe.pb.h>
+#include <caffe/util/io.hpp>
+#include <caffe/vision_layers.hpp>
+#endif
+
 #include "opencv2/viz/vizcore.hpp"
 #include "opencv2/highgui.hpp"
 #include "opencv2/highgui/highgui_c.h"
@@ -132,10 +136,10 @@ class CV_EXPORTS_W IcoSphere
 
 };
 
-class CV_EXPORTS_W Classification
+class CV_EXPORTS_W Feature
 {
 	private:
-		caffe::shared_ptr<caffe::Net<float> > net_;
+		caffe::Net<float>* net_;
 		cv::Size input_geometry_;
 		int num_channels_;
 		cv::Mat mean_;
@@ -150,7 +154,7 @@ class CV_EXPORTS_W Classification
 		/** @brief Convert the input image to the input image format of the network.
 		*/
 	public:
-		Classification();
+		Feature();
 		void list_dir(const char *path,std::vector<string>& files,bool r);
 		/** @brief Get the file name from a root dictionary.
 		*/

modules/cnn_3dobj/samples/classifyIMG_demo.cpp

@@ -32,6 +32,7 @@
  *  POSSIBILITY OF SUCH DAMAGE.
  *
  */
+#define HAVE_CAFFE
 #include <opencv2/cnn_3dobj.hpp>
 #include <iomanip>
 using namespace cv;
@@ -50,7 +51,7 @@ int main(int argc, char** argv)
 		     "{device | CPU | device}"
 		     "{dev_id | 0 | dev_id}";
 	cv::CommandLineParser parser(argc, argv, keys);
-	parser.about("Demo for Sphere View data generation");
+	parser.about("Demo for object data classification and pose estimation");
 	if (parser.has("help"))
 	{
 	  parser.printMessage();

modules/cnn_3dobj/samples/sphereview_3dobj_demo.cpp

@@ -32,6 +32,7 @@
  *  POSSIBILITY OF SUCH DAMAGE.
  *
  */
+#define HAVE_CAFFE
 #include <opencv2/cnn_3dobj.hpp>
 #include <opencv2/viz/vizcore.hpp>
 #include <iostream>

modules/cnn_3dobj/src/cnn_feature.cpp

@@ -6,8 +6,8 @@ namespace cv
 {
 namespace cnn_3dobj
 {
-  Classification::Classification(){};
-  void Classification::list_dir(const char *path,vector<string>& files,bool r)
+  Feature::Feature(){};
+  void Feature::list_dir(const char *path,vector<string>& files,bool r)
   {
     DIR *pDir;
     struct dirent *ent;
@@ -25,7 +25,7 @@ namespace cnn_3dobj
         if(r)
         {
           sprintf(childpath, "%s/%s", path, ent->d_name);
-          Classification::list_dir(childpath,files,false);
+          Feature::list_dir(childpath,files,false);
         }
       }
       else
@@ -36,7 +36,7 @@ namespace cnn_3dobj
     sort(files.begin(),files.end());
   };
 
-  void Classification::NetSetter(const string& model_file, const string& trained_file, const string& mean_file, const string& cpu_only, int device_id)
+  void Feature::NetSetter(const string& model_file, const string& trained_file, const string& mean_file, const string& cpu_only, int device_id)
   {
     if (strcmp(cpu_only.c_str(), "CPU") == 0)
     {
@@ -48,7 +48,7 @@ namespace cnn_3dobj
       caffe::Caffe::SetDevice(device_id);
     }
     /* Load the network. */
-    net_.reset(new Net<float>(model_file, TEST));
+    net_ = new Net<float>(model_file, TEST);
     net_->CopyTrainedLayersFrom(trained_file);
     CHECK_EQ(net_->num_inputs(), 1) << "Network should have exactly one input.";
     CHECK_EQ(net_->num_outputs(), 1) << "Network should have exactly one output.";
@@ -61,14 +61,14 @@ namespace cnn_3dobj
     SetMean(mean_file);
   };
 
-  void Classification::GetLabellist(const std::vector<string>& name_gallery)
+  void Feature::GetLabellist(const std::vector<string>& name_gallery)
   {
     for (unsigned int i = 0; i < name_gallery.size(); ++i)
     labels_.push_back(name_gallery[i]);
   };
 
   /* Return the indices of the top N values of vector v. */
-  std::vector<int> Classification::Argmax(const std::vector<float>& v, int N)
+  std::vector<int> Feature::Argmax(const std::vector<float>& v, int N)
   {
     std::vector<std::pair<float, int> > pairs;
     for (size_t i = 0; i < v.size(); ++i)
@@ -81,7 +81,7 @@ namespace cnn_3dobj
   };
 
   //Return the top N predictions.
-  std::vector<std::pair<string, float> > Classification::Classify(const cv::Mat& reference, const cv::Mat& target, int N)
+  std::vector<std::pair<string, float> > Feature::Classify(const cv::Mat& reference, const cv::Mat& target, int N)
   {
     std::vector<float> output;
     for (int i = 0; i < reference.rows; i++)
@@ -102,7 +102,7 @@ namespace cnn_3dobj
   };
 
   /* Load the mean file in binaryproto format. */
-  void Classification::SetMean(const string& mean_file)
+  void Feature::SetMean(const string& mean_file)
   {
     BlobProto blob_proto;
     ReadProtoFromBinaryFileOrDie(mean_file.c_str(), &blob_proto);
@@ -130,7 +130,7 @@ namespace cnn_3dobj
     mean_ = cv::Mat(input_geometry_, mean.type(), channel_mean);
   };
 
-  void Classification::FeatureExtract(InputArray inputimg, OutputArray feature, bool mean_subtract, std::string featrue_blob)
+  void Feature::FeatureExtract(InputArray inputimg, OutputArray feature, bool mean_subtract, std::string featrue_blob)
   {
     Blob<float>* input_layer = net_->input_blobs()[0];
     input_layer->Reshape(1, num_channels_,
@@ -183,7 +183,7 @@ namespace cnn_3dobj
   * don't need to rely on cudaMemcpy2D. The last preprocessing
   * operation will write the separate channels directly to the input
   * layer. */
-  void Classification::WrapInputLayer(std::vector<cv::Mat>* input_channels)
+  void Feature::WrapInputLayer(std::vector<cv::Mat>* input_channels)
   {
     Blob<float>* input_layer = net_->input_blobs()[0];
     int width = input_layer->width();
@@ -197,7 +197,7 @@ namespace cnn_3dobj
     }
   };
 
-  void Classification::Preprocess(const cv::Mat& img,
+  void Feature::Preprocess(const cv::Mat& img,
 std::vector<cv::Mat>* input_channels, bool mean_subtract)
   {
     /* Convert the input image to the input image format of the network. */

modules/cnn_3dobj/src/precomp.hpp

@@ -42,6 +42,7 @@ the use of this software, even if advised of the possibility of such damage.
 #ifndef __OPENCV_CNN_3DOBJ_PRECOMP_HPP__
 #define __OPENCV_CNN_3DOBJ_PRECOMP_HPP__
 
+#include <opencv2/cnn_3dobj_config.hpp>
 #include <opencv2/cnn_3dobj.hpp>
 
 #endif