diff --git a/samples/CMakeLists.txt b/samples/CMakeLists.txt
index fa7556b70a..f5735e91ea 100644
--- a/samples/CMakeLists.txt
+++ b/samples/CMakeLists.txt
@@ -7,3 +7,4 @@ add_subdirectory(c)
 add_subdirectory(cpp)
 add_subdirectory(swig_python)
 add_subdirectory(octave)
+#add_subdirectory(gpu)
diff --git a/samples/gpu/CMakeLists.txt b/samples/gpu/CMakeLists.txt
new file mode 100644
index 0000000000..c658e8c7da
--- /dev/null
+++ b/samples/gpu/CMakeLists.txt
@@ -0,0 +1,59 @@
+if (BUILD_EXAMPLES)
+    project(gpu_samples)
+
+    include_directories(
+        "${CMAKE_SOURCE_DIR}/modules/core/include"
+        "${CMAKE_SOURCE_DIR}/modules/flann/include"
+        "${CMAKE_SOURCE_DIR}/modules/imgproc/include"
+        "${CMAKE_SOURCE_DIR}/modules/video/include"
+        "${CMAKE_SOURCE_DIR}/modules/highgui/include"
+        "${CMAKE_SOURCE_DIR}/modules/ml/include"
+        "${CMAKE_SOURCE_DIR}/modules/calib3d/include"
+        "${CMAKE_SOURCE_DIR}/modules/features2d/include"
+        "${CMAKE_SOURCE_DIR}/modules/objdetect/include"
+        "${CMAKE_SOURCE_DIR}/modules/legacy/include"
+        "${CMAKE_SOURCE_DIR}/modules/contrib/include"
+        "${CMAKE_SOURCE_DIR}/modules/gpu/include"
+    )   
+
+    if(CMAKE_COMPILER_IS_GNUCXX)
+        set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wno-unused-function")
+    endif()
+    
+    # ---------------------------------------------
+    #      Define executable targets
+    # ---------------------------------------------
+    MACRO(MY_DEFINE_EXAMPLE name srcs)
+        set(the_target "example_${name}")
+    	add_executable(${the_target} ${srcs})
+		set_target_properties(${the_target} PROPERTIES
+		    OUTPUT_NAME "${name}"
+		    PROJECT_LABEL "(EXAMPLE) ${name}")
+    	add_dependencies(${the_target} opencv_core opencv_flann opencv_imgproc opencv_highgui
+    	    opencv_ml opencv_video opencv_objdetect opencv_features2d
+    	    opencv_calib3d opencv_legacy opencv_contrib opencv_gpu)
+    	target_link_libraries(${the_target} ${OPENCV_LINKER_LIBS} opencv_core
+    	    opencv_flann opencv_imgproc opencv_highgui opencv_ml opencv_video opencv_objdetect
+    	    opencv_features2d opencv_calib3d opencv_legacy opencv_contrib opencv_gpu)
+
+        if(WIN32)
+            install(TARGETS ${the_target}
+                RUNTIME DESTINATION "samples/gpu" COMPONENT main)
+        endif()
+    ENDMACRO(MY_DEFINE_EXAMPLE)
+    
+    file(GLOB gpu_samples RELATIVE ${CMAKE_CURRENT_SOURCE_DIR} *.cpp)
+    
+    foreach(sample_filename ${gpu_samples})
+        get_filename_component(sample ${sample_filename} NAME_WE)
+        MY_DEFINE_EXAMPLE(${sample}  ${sample_filename})
+    endforeach()
+endif(BUILD_EXAMPLES)
+
+if (NOT WIN32)
+    file(GLOB GPU_FILES *.c *.cpp *.jpg *.png *.data makefile.* build_all.sh *.dsp *.cmd )
+    install(FILES ${GPU_FILES}
+            DESTINATION share/opencv/samples/gpu
+            PERMISSIONS OWNER_READ GROUP_READ WORLD_READ)
+endif ()
+
diff --git a/samples/gpu/gpu_hog.cpp b/samples/gpu/gpu_hog.cpp
new file mode 100644
index 0000000000..0fd4d94160
--- /dev/null
+++ b/samples/gpu/gpu_hog.cpp
@@ -0,0 +1,373 @@
+#include <iostream>
+#include <fstream>
+#include <string>
+#include <sstream>
+#include <iomanip>
+#include "opencv2/gpu/gpu.hpp"
+#include "opencv2/highgui/highgui.hpp"
+
+using namespace std;
+using namespace cv;
+
+
+/** Contains all properties of application (including those which can be
+changed by user in runtime) */
+class Settings
+{
+public:
+    /** Sets default values */
+    Settings();
+
+    /** Reads settings from command args */
+    static Settings Read(int argc, char** argv);
+
+    string src;
+    bool src_is_video;
+    bool make_gray;
+    bool resize_src;
+    double resize_src_scale;
+    double scale;
+    int nlevels;
+    int gr_threshold;
+    double hit_threshold;
+    int win_width;
+};
+
+
+/** Describes aplication logic */
+class App
+{
+public:    
+    /** Initializes application */
+    App(const Settings& s);
+
+    /** Runs demo using OpenCV highgui module for GUI building */
+    void RunOpencvGui();
+
+    /** Processes user keybord input */
+    void HandleKey(char key);
+
+    void HogWorkBegin();
+    void HogWorkEnd();
+    double HogWorkFps() const;
+
+    void WorkBegin();
+    void WorkEnd();
+    double WorkFps() const;
+
+    const string GetPerformanceSummary() const;
+
+private:
+    App operator=(App&);
+
+    Settings settings;
+    bool running;
+
+    bool use_gpu;
+    bool make_gray;
+    double scale;
+    int gr_threshold;
+    int nlevels;
+    double hit_threshold;
+
+    int64 hog_work_begin;
+    double hog_work_fps;
+
+    int64 work_begin;
+    double work_fps;
+};
+
+
+int main(int argc, char** argv)
+{
+    try
+    {
+        if (argc < 2)
+        {
+            cout << "Usage:\nsample_hog\n" 
+                << "  -src <path_to_the_source>\n"
+                << "  [-src_is_video <true/false>] # says to interp. src as img or as video\n"
+                << "  [-make_gray <true/false>] # convert image to gray one or not\n"
+                << "  [-resize_src <true/false>] # do resize of the source image or not\n"
+                << "  [-resize_src_scale <double>] # preprocessing image scale factor\n"
+                << "  [-hit_threshold <double>] # classifying plane dist. threshold (0.0 usually)\n"
+                << "  [-scale <double>] # HOG window scale factor\n"
+                << "  [-nlevels <int>] # max number of HOG window scales\n"
+                << "  [-win_width <int>] # width of the window (48 or 64)\n"
+                << "  [-gr_threshold <int>] # merging similar rects constant\n";
+            return 1;
+        }
+        App app(Settings::Read(argc, argv));
+        app.RunOpencvGui();
+    }
+    catch (const Exception& e) { return cout << "Error: "  << e.what() << endl, 1; }
+    catch (const exception& e) { return cout << "Error: "  << e.what() << endl, 1; }
+    catch(...) { return cout << "Unknown exception" << endl, 1; }
+    return 0;
+}
+
+
+Settings::Settings()
+{
+    src_is_video = false;
+    make_gray = false;
+    resize_src = true;
+    resize_src_scale = 1.5;
+    scale = 1.05;
+    nlevels = 13;
+    gr_threshold = 8;
+    hit_threshold = 1.4;
+    win_width = 48;
+}
+
+
+Settings Settings::Read(int argc, char** argv)
+{
+    cout << "Parsing command args" << endl;
+
+    Settings settings;
+    for (int i = 1; i < argc - 1; i += 2)
+    {
+        string key = argv[i];
+        string val = argv[i + 1];
+        if (key == "-src") settings.src = val;
+        else if (key == "-src_is_video") settings.src_is_video = (val == "true");
+        else if (key == "-make_gray") settings.make_gray = (val == "true");
+        else if (key == "-resize_src") settings.resize_src = (val == "true");
+        else if (key == "-resize_src_scale") settings.resize_src_scale = atof(val.c_str());
+        else if (key == "-hit_threshold") settings.hit_threshold = atof(val.c_str());
+        else if (key == "-scale") settings.scale = atof(val.c_str());
+        else if (key == "-nlevels") settings.nlevels = atoi(val.c_str());
+        else if (key == "-win_width") settings.win_width = atoi(val.c_str());
+        else if (key == "-gr_threshold") settings.gr_threshold = atoi(val.c_str());
+        else throw exception((string("Unknown key: ") + key).c_str());
+    }
+
+    cout << "Command args are parsed\n";
+    return settings;
+}
+
+
+
+App::App(const Settings &s)
+{
+    settings = s;
+    cout << "\nControls:\n"
+         << "ESC - exit\n"
+         << "m - change mode GPU <-> CPU\n"
+         << "g - convert image to gray or not\n"
+         << "1/q - increase/decrease HOG scale\n"
+         << "2/w - increase/decrease levels count\n"
+         << "3/e - increase/decrease HOG group threshold\n"
+         << "4/r - increase/decrease hit threshold\n"
+         << endl;
+
+    use_gpu = true;
+    make_gray = settings.make_gray;
+    scale = settings.scale;
+    gr_threshold = settings.gr_threshold;
+    nlevels = settings.nlevels;
+    hit_threshold = settings.hit_threshold;
+
+    if (settings.win_width != 64 && settings.win_width != 48)
+        settings.win_width = 64;
+
+    cout << endl << "Scale: " << scale << endl;
+    cout << "Group threshold: " << gr_threshold << endl;
+    cout << "Levels number: " << nlevels << endl;
+    cout << "Win width: " << settings.win_width << endl;
+    cout << "Hit threshold: " << hit_threshold << endl;
+    cout << endl;
+}
+
+void App::RunOpencvGui()
+{
+    running = true;
+
+    Size win_width(settings.win_width, settings.win_width * 2); //(64, 128) or (48, 96)
+
+    vector<float> detector;
+
+    if (win_width == Size(64,128))
+        detector = cv::gpu::HOGDescriptor::getPeopleDetector_64x128();
+    else
+        detector = cv::gpu::HOGDescriptor::getPeopleDetector_48x96();
+
+    // GPU's HOG classifier
+    cv::gpu::HOGDescriptor gpu_hog(win_width);
+    gpu_hog.setSVMDetector(detector);
+
+    // CPU's HOG classifier
+    cv::HOGDescriptor cpu_hog(win_width, Size(16,16), Size(8,8), Size(8,8), 9, 1, -1, HOGDescriptor::L2Hys, 0.2, true, HOGDescriptor::DEFAULT_NLEVELS);
+    cpu_hog.setSVMDetector(detector);
+
+    // Make endless cycle from video (if src is video)
+    while (running)
+    {
+        VideoCapture vc;
+        Mat frame;
+        
+        if (settings.src_is_video)
+        {
+            vc.open(settings.src.c_str());
+            if (!vc.isOpened())
+                throw exception(string("Can't open video file: " + settings.src).c_str());
+            vc >> frame;
+        }
+        else
+            frame = imread(settings.src);
+
+        Mat img_aux, img, img_to_show;
+        gpu::GpuMat gpu_img;
+
+        // Iterate over all frames
+        while (running && !frame.empty())
+        {
+            WorkBegin();
+
+            vector<Rect> found;
+
+            // Change format of the image (input must be 8UC3)
+            if (make_gray)
+                cvtColor(frame, img_aux, CV_BGR2GRAY);
+            else if (use_gpu)
+                cvtColor(frame, img_aux, CV_BGR2BGRA);
+            else
+                img_aux = frame;
+
+            // Resize image
+            if (settings.resize_src)
+                resize(img_aux, img, Size(int(frame.cols * settings.resize_src_scale), int(frame.rows * settings.resize_src_scale)));
+            else
+                img = img_aux;
+            img_to_show = img;
+
+            gpu_hog.nlevels = nlevels;
+            cpu_hog.nlevels = nlevels;
+
+            // Perform HOG classification
+            HogWorkBegin();
+            if (use_gpu)
+            {
+                gpu_img = img;
+                gpu_hog.detectMultiScale(gpu_img, found, hit_threshold, Size(8, 8), Size(0, 0), scale, gr_threshold);
+            }
+            else
+                cpu_hog.detectMultiScale(img, found, hit_threshold, Size(8, 8), Size(0, 0), scale, gr_threshold);
+            HogWorkEnd();
+
+            // Draw positive classified windows
+            for (size_t i = 0; i < found.size(); i++)
+            {
+                Rect r = found[i];
+                rectangle(img_to_show, r.tl(), r.br(), CV_RGB(0, 255, 0), 3);
+            }
+
+            WorkEnd();
+
+            // Show results
+            putText(img_to_show, GetPerformanceSummary(), Point(5, 25), FONT_HERSHEY_SIMPLEX, 1.0, Scalar(0, 0, 255), 2);
+            imshow("opencv_gpu_hog", img_to_show);
+            HandleKey((char)waitKey(1));
+
+            if (settings.src_is_video)
+            {
+                vc >> frame;
+            }
+        }
+    }
+}
+
+
+void App::HandleKey(char key)
+{
+    switch (key)
+    {
+    case 27:
+        running = false;
+        break;
+    case 'm':
+    case 'M':
+        use_gpu = !use_gpu;
+        cout << "Switched to " << (use_gpu ? "CUDA" : "CPU") << " mode\n";
+        break;
+    case 'g':
+    case 'G':
+        make_gray = !make_gray;
+        cout << "Convert image to gray: " << (make_gray ? "YES" : "NO") << endl;
+        break;
+    case '1':
+        scale *= 1.05;
+        cout << "Scale: " << scale << endl;
+        break;
+    case 'q':
+    case 'Q':
+        scale /= 1.05;
+        cout << "Scale: " << scale << endl;
+        break;
+    case '2':
+        nlevels++;
+        cout << "Levels number: " << nlevels << endl;
+        break;
+    case 'w':
+    case 'W':
+        nlevels = max(nlevels - 1, 1);
+        cout << "Levels number: " << nlevels << endl;
+        break;
+    case '3':
+        gr_threshold++;
+        cout << "Group threshold: " << gr_threshold << endl;
+        break;
+    case 'e':
+    case 'E':
+        gr_threshold = max(0, gr_threshold - 1);
+        cout << "Group threshold: " << gr_threshold << endl;
+        break;
+    case '4':
+        hit_threshold+=0.25;
+        cout << "Hit threshold: " << hit_threshold << endl;
+        break;
+    case 'r':
+    case 'R':
+        hit_threshold = max(0.0, hit_threshold - 0.25);
+        cout << "Hit threshold: " << hit_threshold << endl;
+        break;
+    }
+}
+
+
+inline void App::HogWorkBegin() { hog_work_begin = getTickCount(); }
+
+
+inline void App::HogWorkEnd() 
+{
+    int64 delta = getTickCount() - hog_work_begin;
+    double freq = getTickFrequency();
+    hog_work_fps = freq / delta;
+}
+
+
+inline double App::HogWorkFps() const { return hog_work_fps; }
+
+
+inline void App::WorkBegin() { work_begin = getTickCount(); }
+
+
+inline void App::WorkEnd() 
+{
+    int64 delta = getTickCount() - work_begin;
+    double freq = getTickFrequency();
+    work_fps = freq / delta;
+}
+
+
+inline double App::WorkFps() const { return work_fps; }
+
+
+inline const string App::GetPerformanceSummary() const 
+{ 
+    stringstream ss;
+    ss << (use_gpu ? "GPU" : "CPU") << " HOG FPS: " << setiosflags(ios::left) << setprecision(4) <<
+       setw(7) << HogWorkFps() << " Total FPS: " << setprecision(4) << setw(7) << WorkFps();
+    return ss.str();
+}
\ No newline at end of file
diff --git a/samples/gpu/road.png b/samples/gpu/road.png
new file mode 100644
index 0000000000..a52c87cffb
Binary files /dev/null and b/samples/gpu/road.png differ