diff --git a/modules/gpu/CMakeLists.txt b/modules/gpu/CMakeLists.txt
index 887daa917..1227b93a3 100644
--- a/modules/gpu/CMakeLists.txt
+++ b/modules/gpu/CMakeLists.txt
@@ -5,7 +5,7 @@ endif()
 set(the_description "GPU-accelerated Computer Vision")
 ocv_add_module(gpu opencv_imgproc opencv_calib3d opencv_objdetect opencv_video opencv_nonfree opencv_photo opencv_legacy)
 
-ocv_module_include_directories("${CMAKE_CURRENT_SOURCE_DIR}/src/cuda" "${OpenCV_SOURCE_DIR}/modules/highgui/src")
+ocv_module_include_directories("${CMAKE_CURRENT_SOURCE_DIR}/src/cuda" "${CMAKE_CURRENT_SOURCE_DIR}/../highgui/src")
 
 file(GLOB lib_hdrs               "include/opencv2/${name}/*.hpp"               "include/opencv2/${name}/*.h")
 file(GLOB lib_device_hdrs        "include/opencv2/${name}/device/*.hpp"        "include/opencv2/${name}/device/*.h")
@@ -21,16 +21,16 @@ source_group("Src\\Cuda"      FILES ${lib_cuda} ${lib_cuda_hdrs})
 source_group("Device"         FILES ${lib_device_hdrs})
 source_group("Device\\Detail" FILES ${lib_device_hdrs_detail})
 
-if(HAVE_CUDA)
+if (HAVE_CUDA)
   file(GLOB_RECURSE ncv_srcs "src/nvidia/*.cpp" "src/nvidia/*.h*")
-  file(GLOB_RECURSE ncv_cuda "src/nvidia/*.cu")
+  file(GLOB_RECURSE ncv_cuda "src/nvidia/*.cu")  
   set(ncv_files ${ncv_srcs} ${ncv_cuda})
 
   source_group("Src\\NVidia" FILES ${ncv_files})
-
-  ocv_include_directories("${CMAKE_CURRENT_SOURCE_DIR}/src/nvidia" "${CMAKE_CURRENT_SOURCE_DIR}/src/nvidia/core" "${CMAKE_CURRENT_SOURCE_DIR}/src/nvidia/NPP_staging" ${CUDA_INCLUDE_DIRS})
+  ocv_include_directories("src/nvidia" "src/nvidia/core" "src/nvidia/NPP_staging" ${CUDA_INCLUDE_DIRS})
   ocv_warnings_disable(CMAKE_CXX_FLAGS -Wundef -Wmissing-declarations /wd4211 /wd4201 /wd4100 /wd4505 /wd4408)
-  string(REPLACE "-Wsign-promo" "" CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
+  string(REPLACE "-Wsign-promo" "" CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")  
+  #set (CUDA_NVCC_FLAGS ${CUDA_NVCC_FLAGS} "-Xcompiler;/EHsc-;")
 
   if(MSVC)
     if(NOT ENABLE_NOISY_WARNINGS)
@@ -46,13 +46,21 @@ if(HAVE_CUDA)
 
   set(cuda_link_libs ${CUDA_LIBRARIES} ${CUDA_npp_LIBRARY})
 
-  if(WIN32)
+  if(NOT APPLE)
+    unset(CUDA_nvcuvid_LIBRARY CACHE)
     find_cuda_helper_libs(nvcuvid)
     set(cuda_link_libs ${cuda_link_libs} ${CUDA_nvcuvid_LIBRARY})
+  endif()
 
+  if(WIN32)
+    unset(CUDA_nvcuvenc_LIBRARY CACHE)
     find_cuda_helper_libs(nvcuvenc)
     set(cuda_link_libs ${cuda_link_libs} ${CUDA_nvcuvenc_LIBRARY})
   endif()
+
+  if(NOT APPLE AND WITH_FFMPEG)
+    set(cuda_link_libs ${cuda_link_libs} ${HIGHGUI_LIBRARIES})
+  endif()
 else()
   set(lib_cuda "")
   set(cuda_objs "")
diff --git a/modules/gpu/perf/perf_video.cpp b/modules/gpu/perf/perf_video.cpp
index 5afe0d15f..bf2fd99c6 100644
--- a/modules/gpu/perf/perf_video.cpp
+++ b/modules/gpu/perf/perf_video.cpp
@@ -1045,12 +1045,10 @@ PERF_TEST_P(Video_Cn_MaxFeatures, DISABLED_Video_GMG,
     }
 }
 
-#ifdef WIN32
-
 //////////////////////////////////////////////////////
 // VideoWriter
 
-PERF_TEST_P(Video, Video_VideoWriter, Values("gpu/video/768x576.avi", "gpu/video/1920x1080.avi"))
+PERF_TEST_P(Video, DISABLED_Video_VideoWriter, Values("gpu/video/768x576.avi", "gpu/video/1920x1080.avi"))
 {
     declare.time(30);
 
@@ -1147,6 +1145,4 @@ PERF_TEST_P(Video, Video_VideoReader, Values("gpu/video/768x576.avi", "gpu/video
     }
 }
 
-#endif
-
 } // namespace
diff --git a/modules/gpu/src/cuvid_video_source.cpp b/modules/gpu/src/cuvid_video_source.cpp
index 0700ba4cd..5c4691353 100644
--- a/modules/gpu/src/cuvid_video_source.cpp
+++ b/modules/gpu/src/cuvid_video_source.cpp
@@ -1,7 +1,7 @@
 #include "cuvid_video_source.h"
 #include "cu_safe_call.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 cv::gpu::detail::CuvidVideoSource::CuvidVideoSource(const std::string& fname)
 {
diff --git a/modules/gpu/src/cuvid_video_source.h b/modules/gpu/src/cuvid_video_source.h
index b9c522cae..a06353fed 100644
--- a/modules/gpu/src/cuvid_video_source.h
+++ b/modules/gpu/src/cuvid_video_source.h
@@ -45,7 +45,7 @@
 
 #include "precomp.hpp"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 namespace cv { namespace gpu
 {
diff --git a/modules/gpu/src/ffmpeg_video_source.cpp b/modules/gpu/src/ffmpeg_video_source.cpp
index fe9771554..7538de601 100644
--- a/modules/gpu/src/ffmpeg_video_source.cpp
+++ b/modules/gpu/src/ffmpeg_video_source.cpp
@@ -42,7 +42,7 @@
 
 #include "ffmpeg_video_source.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 #if defined(HAVE_FFMPEG) && defined(BUILD_SHARED_LIBS)
     #include "cap_ffmpeg_impl.hpp"
diff --git a/modules/gpu/src/ffmpeg_video_source.h b/modules/gpu/src/ffmpeg_video_source.h
index 9ec5559c5..66bb7c71c 100644
--- a/modules/gpu/src/ffmpeg_video_source.h
+++ b/modules/gpu/src/ffmpeg_video_source.h
@@ -46,7 +46,7 @@
 #include "precomp.hpp"
 #include "thread_wrappers.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 struct InputMediaStream_FFMPEG;
 
diff --git a/modules/gpu/src/frame_queue.cpp b/modules/gpu/src/frame_queue.cpp
index 5f0659c7b..4ed6d0f14 100644
--- a/modules/gpu/src/frame_queue.cpp
+++ b/modules/gpu/src/frame_queue.cpp
@@ -42,7 +42,7 @@
 
 #include "frame_queue.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 cv::gpu::detail::FrameQueue::FrameQueue() :
     endOfDecode_(0),
diff --git a/modules/gpu/src/frame_queue.h b/modules/gpu/src/frame_queue.h
index dc3da7336..5c049a28e 100644
--- a/modules/gpu/src/frame_queue.h
+++ b/modules/gpu/src/frame_queue.h
@@ -46,7 +46,7 @@
 #include "precomp.hpp"
 #include "thread_wrappers.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 namespace cv { namespace gpu
 {
diff --git a/modules/gpu/src/precomp.hpp b/modules/gpu/src/precomp.hpp
index 9310e69f1..c37332ae4 100644
--- a/modules/gpu/src/precomp.hpp
+++ b/modules/gpu/src/precomp.hpp
@@ -97,8 +97,11 @@
         #include <cublas.h>
     #endif
 
-    #ifdef WIN32
+    #ifndef __APPLE__
         #include <nvcuvid.h>
+    #endif
+
+    #ifdef WIN32
         #include <NVEncoderAPI.h>
     #endif
 
diff --git a/modules/gpu/src/thread_wrappers.cpp b/modules/gpu/src/thread_wrappers.cpp
index e88b79243..fcfa59827 100644
--- a/modules/gpu/src/thread_wrappers.cpp
+++ b/modules/gpu/src/thread_wrappers.cpp
@@ -42,7 +42,7 @@
 
 #include "thread_wrappers.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 #ifdef WIN32
     #define NOMINMAX
diff --git a/modules/gpu/src/thread_wrappers.h b/modules/gpu/src/thread_wrappers.h
index 6d6e87c5e..9198d3e90 100644
--- a/modules/gpu/src/thread_wrappers.h
+++ b/modules/gpu/src/thread_wrappers.h
@@ -45,7 +45,7 @@
 
 #include "precomp.hpp"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 namespace cv { namespace gpu
 {
diff --git a/modules/gpu/src/video_decoder.cpp b/modules/gpu/src/video_decoder.cpp
index bba28af86..41e1d2607 100644
--- a/modules/gpu/src/video_decoder.cpp
+++ b/modules/gpu/src/video_decoder.cpp
@@ -43,7 +43,7 @@
 #include "video_decoder.h"
 #include "frame_queue.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 void cv::gpu::detail::VideoDecoder::create(const VideoReader_GPU::FormatInfo& videoFormat)
 {
diff --git a/modules/gpu/src/video_decoder.h b/modules/gpu/src/video_decoder.h
index 0b638a0c6..9c4dfa929 100644
--- a/modules/gpu/src/video_decoder.h
+++ b/modules/gpu/src/video_decoder.h
@@ -46,7 +46,7 @@
 #include "precomp.hpp"
 #include "cu_safe_call.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 namespace cv { namespace gpu
 {
diff --git a/modules/gpu/src/video_parser.cpp b/modules/gpu/src/video_parser.cpp
index 78f1efe82..2ff14bd75 100644
--- a/modules/gpu/src/video_parser.cpp
+++ b/modules/gpu/src/video_parser.cpp
@@ -43,7 +43,7 @@
 #include "video_parser.h"
 #include "cu_safe_call.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 cv::gpu::detail::VideoParser::VideoParser(VideoDecoder* videoDecoder, FrameQueue* frameQueue) :
     videoDecoder_(videoDecoder), frameQueue_(frameQueue), unparsedPackets_(0), hasError_(false)
diff --git a/modules/gpu/src/video_parser.h b/modules/gpu/src/video_parser.h
index c3c2d1c61..3935ffcef 100644
--- a/modules/gpu/src/video_parser.h
+++ b/modules/gpu/src/video_parser.h
@@ -48,7 +48,7 @@
 #include "frame_queue.h"
 #include "video_decoder.h"
 
-#if defined(HAVE_CUDA) && defined(WIN32)
+#if defined(HAVE_CUDA) && !defined(__APPLE__)
 
 namespace cv { namespace gpu
 {
diff --git a/modules/gpu/src/video_reader.cpp b/modules/gpu/src/video_reader.cpp
index d9294d861..df3c2ff2f 100644
--- a/modules/gpu/src/video_reader.cpp
+++ b/modules/gpu/src/video_reader.cpp
@@ -42,7 +42,7 @@
 
 #include "precomp.hpp"
 
-#if !defined HAVE_CUDA || defined(CUDA_DISABLER) || !defined(WIN32)
+#if !defined HAVE_CUDA || defined(CUDA_DISABLER) || defined(__APPLE__)
 
 class cv::gpu::VideoReader_GPU::Impl
 {
diff --git a/modules/gpu/test/test_video.cpp b/modules/gpu/test/test_video.cpp
index 4cbbf0bf8..424bdbd3b 100644
--- a/modules/gpu/test/test_video.cpp
+++ b/modules/gpu/test/test_video.cpp
@@ -43,11 +43,11 @@
 
 #ifdef HAVE_CUDA
 
-#ifdef WIN32
-
 //////////////////////////////////////////////////////
 // VideoWriter
 
+#ifdef WIN32
+
 PARAM_TEST_CASE(VideoWriter, cv::gpu::DeviceInfo, std::string)
 {
     cv::gpu::DeviceInfo devInfo;
@@ -107,9 +107,13 @@ INSTANTIATE_TEST_CASE_P(GPU_Video, VideoWriter, testing::Combine(
     ALL_DEVICES,
     testing::Values(std::string("768x576.avi"), std::string("1920x1080.avi"))));
 
+#endif // WIN32
+
 //////////////////////////////////////////////////////
 // VideoReader
 
+#ifdef HAVE_NVCUVID
+
 PARAM_TEST_CASE(VideoReader, cv::gpu::DeviceInfo, std::string)
 {
     cv::gpu::DeviceInfo devInfo;
@@ -147,6 +151,6 @@ INSTANTIATE_TEST_CASE_P(GPU_Video, VideoReader, testing::Combine(
     ALL_DEVICES,
     testing::Values(std::string("768x576.avi"), std::string("1920x1080.avi"))));
 
-#endif // WIN32
+#endif // HAVE_NVCUVID
 
 #endif // HAVE_CUDA