avoid calling of setNumThreads() to respect user settings

- bgfg_segm
- peopledetect
- opencv_version
- dnn/colorization
- tapi/opencl_custom_kernel
- tapi/dense_optical_flow (renamed tvl1_optical_flow)

3rdparty/cpufeatures/CMakeLists.txt

@@ -31,3 +31,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${OPENCV_CPUFEATURES_TARGET_NAME} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(cpufeatures LICENSE README.md)

3rdparty/include/opencl/LICENSE.txt

@@ -0,0 +1,25 @@
+Copyright (c) 2008-2015 The Khronos Group Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a
+copy of this software and/or associated documentation files (the
+"Materials"), to deal in the Materials without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Materials, and to
+permit persons to whom the Materials are furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be included
+in all copies or substantial portions of the Materials.
+
+MODIFICATIONS TO THIS FILE MAY MEAN IT NO LONGER ACCURATELY REFLECTS
+KHRONOS STANDARDS. THE UNMODIFIED, NORMATIVE VERSIONS OF KHRONOS
+SPECIFICATIONS AND HEADER INFORMATION ARE LOCATED AT
+   https://www.khronos.org/registry/
+
+THE MATERIALS ARE PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+MATERIALS OR THE USE OR OTHER DEALINGS IN THE MATERIALS.

3rdparty/ittnotify/CMakeLists.txt

@@ -60,3 +60,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${ITT_LIBRARY} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(ittnotify src/ittnotify/LICENSE.BSD src/ittnotify/LICENSE.GPL)

3rdparty/libjasper/CMakeLists.txt

@@ -47,3 +47,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${JASPER_LIBRARY} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(jasper LICENSE README copyright)

3rdparty/libjpeg/CMakeLists.txt

@@ -51,3 +51,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${JPEG_LIBRARY} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(libjpeg README)

3rdparty/libpng/CMakeLists.txt

@@ -48,7 +48,7 @@ if(PPC64LE OR PPC64)
     list(APPEND lib_srcs powerpc/powerpc_init.c powerpc/filter_vsx_intrinsics.c)
     add_definitions(-DPNG_POWERPC_VSX_OPT=2)
   else()
-    add_definitions(-DPNG_POWERPC_VSX_OPT=0)    
+    add_definitions(-DPNG_POWERPC_VSX_OPT=0)
   endif()
 endif()
 
@@ -80,3 +80,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${PNG_LIBRARY} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(libpng LICENSE README opencv-libpng.patch)

3rdparty/libtiff/CMakeLists.txt

@@ -470,3 +470,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${TIFF_LIBRARY} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(libtiff COPYRIGHT)

3rdparty/openexr/CMakeLists.txt

@@ -76,5 +76,7 @@ if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(IlmImf EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
 
+ocv_install_3rdparty_licenses(openexr LICENSE AUTHORS.ilmbase AUTHORS.openexr fix_msvc2013_errors.patch)
+
 set(OPENEXR_INCLUDE_PATHS ${OPENEXR_INCLUDE_PATHS} PARENT_SCOPE)
 set(OPENEXR_VERSION "1.7.1" PARENT_SCOPE)

3rdparty/protobuf/CMakeLists.txt

@@ -150,3 +150,5 @@ set(Protobuf_VERSION ${Protobuf_VERSION} CACHE INTERNAL "" FORCE)
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(libprotobuf EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(protobuf LICENSE README.md)

3rdparty/protobuf/README.md

@@ -1,3 +1,3 @@
 Project: Protocol Buffers - Google's data interchange format
 Source code: https://github.com/google/protobuf
-Version: 3.1.0
+Version: 3.5.1

3rdparty/tbb/CMakeLists.txt

@@ -156,4 +156,6 @@ ocv_install_target(tbb EXPORT OpenCVModules
     ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev
     )
 
+ocv_install_3rdparty_licenses(tbb "${tbb_src_dir}/LICENSE" "${tbb_src_dir}/README")
+
 ocv_tbb_read_version("${tbb_src_dir}/include")

3rdparty/zlib/CMakeLists.txt

@@ -99,3 +99,5 @@ endif()
 if(NOT BUILD_SHARED_LIBS)
   ocv_install_target(${ZLIB_LIBRARY} EXPORT OpenCVModules ARCHIVE DESTINATION ${OPENCV_3P_LIB_INSTALL_PATH} COMPONENT dev)
 endif()
+
+ocv_install_3rdparty_licenses(zlib README)

cmake/OpenCVDetectOpenCL.cmake

@@ -5,6 +5,7 @@ if(APPLE)
 else(APPLE)
   set(OPENCL_LIBRARY "" CACHE STRING "OpenCL library")
   set(OPENCL_INCLUDE_DIR "${OpenCV_SOURCE_DIR}/3rdparty/include/opencl/1.2" CACHE PATH "OpenCL include directory")
+  ocv_install_3rdparty_licenses(opencl-headers "${OpenCV_SOURCE_DIR}/3rdparty/include/opencl/LICENSE.txt")
 endif(APPLE)
 mark_as_advanced(OPENCL_INCLUDE_DIR OPENCL_LIBRARY)
 

cmake/OpenCVFindIPP.cmake

@@ -244,6 +244,7 @@ if(NOT DEFINED IPPROOT)
   if(NOT IPPROOT)
     return()
   endif()
+  ocv_install_3rdparty_licenses(ippicv "${IPPROOT}/readme.htm" "${IPPROOT}/license/ippEULA.txt")
 endif()
 
 file(TO_CMAKE_PATH "${IPPROOT}" __IPPROOT)

cmake/OpenCVFindIPPIW.cmake

@@ -136,6 +136,13 @@ if(BUILD_IPP_IW)
   # local sources
   ippiw_setup("${OpenCV_SOURCE_DIR}/3rdparty/ippiw" 1)
 
+  set(IPPIW_ROOT "${IPPROOT}/../${IW_PACKAGE_SUBDIR}")
+  ocv_install_3rdparty_licenses(ippiw
+    "${IPPIW_ROOT}/EULA.txt"
+    "${IPPIW_ROOT}/redist.txt"
+    "${IPPIW_ROOT}/support.txt"
+    "${IPPIW_ROOT}/third-party-programs.txt")
+
   # Package sources
   get_filename_component(__PATH "${IPPROOT}/../${IW_PACKAGE_SUBDIR}/" ABSOLUTE)
   ippiw_setup("${__PATH}" 1)
@@ -161,9 +168,15 @@ if(NOT HAVE_IPP_ICV AND BUILD_IPP_IW)
   set(TEMP_ROOT 0)
   include("${OpenCV_SOURCE_DIR}/3rdparty/ippicv/ippicv.cmake")
   download_ippicv(TEMP_ROOT)
+  set(IPPIW_ROOT "${TEMP_ROOT}/../${IW_PACKAGE_SUBDIR}")
+  ocv_install_3rdparty_licenses(ippiw
+    "${IPPIW_ROOT}/EULA.txt"
+    "${IPPIW_ROOT}/redist.txt"
+    "${IPPIW_ROOT}/support.txt"
+    "${IPPIW_ROOT}/third-party-programs.txt")
 
   # Package sources. Only sources are compatible with regular Intel IPP
-  ippiw_setup("${TEMP_ROOT}/../${IW_PACKAGE_SUBDIR}/" 1)
+  ippiw_setup("${IPPIW_ROOT}" 1)
 endif()
 
 

cmake/OpenCVModule.cmake

@@ -747,7 +747,7 @@ endmacro()
 
 # finds and sets headers and sources for the standard OpenCV module
 # Usage:
-# ocv_glob_module_sources([EXCLUDE_CUDA] <extra sources&headers in the same format as used in ocv_set_module_sources>)
+# ocv_glob_module_sources([EXCLUDE_CUDA] [EXCLUDE_OPENCL] <extra sources&headers in the same format as used in ocv_set_module_sources>)
 macro(ocv_glob_module_sources)
   ocv_debug_message("ocv_glob_module_sources(" ${ARGN} ")")
   set(_argn ${ARGN})
@@ -755,6 +755,10 @@ macro(ocv_glob_module_sources)
   if(NOT exclude_cuda EQUAL -1)
     list(REMOVE_AT _argn ${exclude_cuda})
   endif()
+  list(FIND _argn "EXCLUDE_OPENCL" exclude_opencl)
+  if(NOT exclude_opencl EQUAL -1)
+    list(REMOVE_AT _argn ${exclude_opencl})
+  endif()
 
   file(GLOB_RECURSE lib_srcs
        "${CMAKE_CURRENT_LIST_DIR}/src/*.cpp"
@@ -801,7 +805,7 @@ macro(ocv_glob_module_sources)
   file(GLOB cl_kernels
        "${CMAKE_CURRENT_LIST_DIR}/src/opencl/*.cl"
   )
-  if(cl_kernels)
+  if(cl_kernels AND exclude_opencl EQUAL -1)
     set(OCL_NAME opencl_kernels_${name})
     add_custom_command(
       OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/${OCL_NAME}.cpp"  # don't add .hpp file here to optimize build process

cmake/OpenCVUtils.cmake

@@ -1002,6 +1002,18 @@ function(ocv_install_target)
   endif()
 endfunction()
 
+# ocv_install_3rdparty_licenses(<library-name> <filename1> [<filename2> ..])
+function(ocv_install_3rdparty_licenses library)
+  foreach(filename ${ARGN})
+    get_filename_component(name "${filename}" NAME)
+    install(
+      FILES "${filename}"
+      DESTINATION "${OPENCV_OTHER_INSTALL_PATH}/licenses"
+      COMPONENT licenses
+      RENAME "${library}-${name}"
+      OPTIONAL)
+  endforeach()
+endfunction()
 
 # read set of version defines from the header file
 macro(ocv_parse_header FILENAME FILE_VAR)

modules/calib3d/perf/perf_stereosgbm.cpp

@@ -67,7 +67,6 @@ PERF_TEST_P( TestStereoCorresp, DISABLED_TooLongInDebug_SGBM, Combine(Values(Siz
 
     MakeArtificialExample(rng,src_left,src_right);
 
-    cv::setNumThreads(cv::getNumberOfCPUs());
     int wsize = 3;
     int P1 = 8*src_left.channels()*wsize*wsize;
     TEST_CYCLE()

modules/calib3d/test/test_solvepnp_ransac.cpp

@@ -382,6 +382,7 @@ TEST(Calib3d_SolvePnPRansac, concurrency)
     Mat rvec1, rvec2;
     Mat tvec1, tvec2;
 
+    int threads = getNumThreads();
     {
         // limit concurrency to get deterministic result
         theRNG().state = 20121010;
@@ -390,6 +391,7 @@ TEST(Calib3d_SolvePnPRansac, concurrency)
     }
 
     {
+        setNumThreads(threads);
         Mat rvec;
         Mat tvec;
         // parallel executions

modules/core/3rdparty/SoftFloat/COPYING.txt

@@ -0,0 +1,36 @@
+
+License for Berkeley SoftFloat Release 3c
+
+John R. Hauser
+2017 February 10
+
+The following applies to the whole of SoftFloat Release 3c as well as to
+each source file individually.
+
+Copyright 2011, 2012, 2013, 2014, 2015, 2016, 2017 The Regents of the
+University of California.  All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+ 1. Redistributions of source code must retain the above copyright notice,
+    this list of conditions, and the following disclaimer.
+
+ 2. Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions, and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+ 3. Neither the name of the University nor the names of its contributors
+    may be used to endorse or promote products derived from this software
+    without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS", AND ANY
+EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, ARE
+DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY
+DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

modules/core/CMakeLists.txt

@@ -70,3 +70,5 @@ endif()
 
 ocv_add_accuracy_tests()
 ocv_add_perf_tests()
+
+ocv_install_3rdparty_licenses(SoftFloat "${CMAKE_CURRENT_SOURCE_DIR}/3rdparty/SoftFloat/COPYING.txt")

modules/core/include/opencv2/core/version.hpp

@@ -8,7 +8,7 @@
 #define CV_VERSION_MAJOR    3
 #define CV_VERSION_MINOR    4
 #define CV_VERSION_REVISION 1
-#define CV_VERSION_STATUS   ""
+#define CV_VERSION_STATUS   "-cvsdk"
 
 #define CVAUX_STR_EXP(__A)  #__A
 #define CVAUX_STR(__A)      CVAUX_STR_EXP(__A)

modules/core/src/ocl.cpp

@@ -6015,98 +6015,85 @@ const char* convertTypeStr(int sdepth, int ddepth, int cn, char* buf)
 
 const char* getOpenCLErrorString(int errorCode)
 {
+#define CV_OCL_CODE(id) case id: return #id
+#define CV_OCL_CODE_(id, name) case id: return #name
     switch (errorCode)
     {
-    case   0: return "CL_SUCCESS";
+    CV_OCL_CODE(CL_SUCCESS);
-    case  -1: return "CL_DEVICE_NOT_FOUND";
+    CV_OCL_CODE(CL_DEVICE_NOT_FOUND);
-    case  -2: return "CL_DEVICE_NOT_AVAILABLE";
+    CV_OCL_CODE(CL_DEVICE_NOT_AVAILABLE);
-    case  -3: return "CL_COMPILER_NOT_AVAILABLE";
+    CV_OCL_CODE(CL_COMPILER_NOT_AVAILABLE);
-    case  -4: return "CL_MEM_OBJECT_ALLOCATION_FAILURE";
+    CV_OCL_CODE(CL_MEM_OBJECT_ALLOCATION_FAILURE);
-    case  -5: return "CL_OUT_OF_RESOURCES";
+    CV_OCL_CODE(CL_OUT_OF_RESOURCES);
-    case  -6: return "CL_OUT_OF_HOST_MEMORY";
+    CV_OCL_CODE(CL_OUT_OF_HOST_MEMORY);
-    case  -7: return "CL_PROFILING_INFO_NOT_AVAILABLE";
+    CV_OCL_CODE(CL_PROFILING_INFO_NOT_AVAILABLE);
-    case  -8: return "CL_MEM_COPY_OVERLAP";
+    CV_OCL_CODE(CL_MEM_COPY_OVERLAP);
-    case  -9: return "CL_IMAGE_FORMAT_MISMATCH";
+    CV_OCL_CODE(CL_IMAGE_FORMAT_MISMATCH);
-    case -10: return "CL_IMAGE_FORMAT_NOT_SUPPORTED";
+    CV_OCL_CODE(CL_IMAGE_FORMAT_NOT_SUPPORTED);
-    case -11: return "CL_BUILD_PROGRAM_FAILURE";
+    CV_OCL_CODE(CL_BUILD_PROGRAM_FAILURE);
-    case -12: return "CL_MAP_FAILURE";
+    CV_OCL_CODE(CL_MAP_FAILURE);
-    case -13: return "CL_MISALIGNED_SUB_BUFFER_OFFSET";
+    CV_OCL_CODE(CL_MISALIGNED_SUB_BUFFER_OFFSET);
-    case -14: return "CL_EXEC_STATUS_ERROR_FOR_EVENTS_IN_WAIT_LIST";
+    CV_OCL_CODE(CL_EXEC_STATUS_ERROR_FOR_EVENTS_IN_WAIT_LIST);
-    case -15: return "CL_COMPILE_PROGRAM_FAILURE";
+    CV_OCL_CODE(CL_COMPILE_PROGRAM_FAILURE);
-    case -16: return "CL_LINKER_NOT_AVAILABLE";
+    CV_OCL_CODE(CL_LINKER_NOT_AVAILABLE);
-    case -17: return "CL_LINK_PROGRAM_FAILURE";
+    CV_OCL_CODE(CL_LINK_PROGRAM_FAILURE);
-    case -18: return "CL_DEVICE_PARTITION_FAILED";
+    CV_OCL_CODE(CL_DEVICE_PARTITION_FAILED);
-    case -19: return "CL_KERNEL_ARG_INFO_NOT_AVAILABLE";
+    CV_OCL_CODE(CL_KERNEL_ARG_INFO_NOT_AVAILABLE);
-    case -30: return "CL_INVALID_VALUE";
+    CV_OCL_CODE(CL_INVALID_VALUE);
-    case -31: return "CL_INVALID_DEVICE_TYPE";
+    CV_OCL_CODE(CL_INVALID_DEVICE_TYPE);
-    case -32: return "CL_INVALID_PLATFORM";
+    CV_OCL_CODE(CL_INVALID_PLATFORM);
-    case -33: return "CL_INVALID_DEVICE";
+    CV_OCL_CODE(CL_INVALID_DEVICE);
-    case -34: return "CL_INVALID_CONTEXT";
+    CV_OCL_CODE(CL_INVALID_CONTEXT);
-    case -35: return "CL_INVALID_QUEUE_PROPERTIES";
+    CV_OCL_CODE(CL_INVALID_QUEUE_PROPERTIES);
-    case -36: return "CL_INVALID_COMMAND_QUEUE";
+    CV_OCL_CODE(CL_INVALID_COMMAND_QUEUE);
-    case -37: return "CL_INVALID_HOST_PTR";
+    CV_OCL_CODE(CL_INVALID_HOST_PTR);
-    case -38: return "CL_INVALID_MEM_OBJECT";
+    CV_OCL_CODE(CL_INVALID_MEM_OBJECT);
-    case -39: return "CL_INVALID_IMAGE_FORMAT_DESCRIPTOR";
+    CV_OCL_CODE(CL_INVALID_IMAGE_FORMAT_DESCRIPTOR);
-    case -40: return "CL_INVALID_IMAGE_SIZE";
+    CV_OCL_CODE(CL_INVALID_IMAGE_SIZE);
-    case -41: return "CL_INVALID_SAMPLER";
+    CV_OCL_CODE(CL_INVALID_SAMPLER);
-    case -42: return "CL_INVALID_BINARY";
+    CV_OCL_CODE(CL_INVALID_BINARY);
-    case -43: return "CL_INVALID_BUILD_OPTIONS";
+    CV_OCL_CODE(CL_INVALID_BUILD_OPTIONS);
-    case -44: return "CL_INVALID_PROGRAM";
+    CV_OCL_CODE(CL_INVALID_PROGRAM);
-    case -45: return "CL_INVALID_PROGRAM_EXECUTABLE";
+    CV_OCL_CODE(CL_INVALID_PROGRAM_EXECUTABLE);
-    case -46: return "CL_INVALID_KERNEL_NAME";
+    CV_OCL_CODE(CL_INVALID_KERNEL_NAME);
-    case -47: return "CL_INVALID_KERNEL_DEFINITION";
+    CV_OCL_CODE(CL_INVALID_KERNEL_DEFINITION);
-    case -48: return "CL_INVALID_KERNEL";
+    CV_OCL_CODE(CL_INVALID_KERNEL);
-    case -49: return "CL_INVALID_ARG_INDEX";
+    CV_OCL_CODE(CL_INVALID_ARG_INDEX);
-    case -50: return "CL_INVALID_ARG_VALUE";
+    CV_OCL_CODE(CL_INVALID_ARG_VALUE);
-    case -51: return "CL_INVALID_ARG_SIZE";
+    CV_OCL_CODE(CL_INVALID_ARG_SIZE);
-    case -52: return "CL_INVALID_KERNEL_ARGS";
+    CV_OCL_CODE(CL_INVALID_KERNEL_ARGS);
-    case -53: return "CL_INVALID_WORK_DIMENSION";
+    CV_OCL_CODE(CL_INVALID_WORK_DIMENSION);
-    case -54: return "CL_INVALID_WORK_GROUP_SIZE";
+    CV_OCL_CODE(CL_INVALID_WORK_GROUP_SIZE);
-    case -55: return "CL_INVALID_WORK_ITEM_SIZE";
+    CV_OCL_CODE(CL_INVALID_WORK_ITEM_SIZE);
-    case -56: return "CL_INVALID_GLOBAL_OFFSET";
+    CV_OCL_CODE(CL_INVALID_GLOBAL_OFFSET);
-    case -57: return "CL_INVALID_EVENT_WAIT_LIST";
+    CV_OCL_CODE(CL_INVALID_EVENT_WAIT_LIST);
-    case -58: return "CL_INVALID_EVENT";
+    CV_OCL_CODE(CL_INVALID_EVENT);
-    case -59: return "CL_INVALID_OPERATION";
+    CV_OCL_CODE(CL_INVALID_OPERATION);
-    case -60: return "CL_INVALID_GL_OBJECT";
+    CV_OCL_CODE(CL_INVALID_GL_OBJECT);
-    case -61: return "CL_INVALID_BUFFER_SIZE";
+    CV_OCL_CODE(CL_INVALID_BUFFER_SIZE);
-    case -62: return "CL_INVALID_MIP_LEVEL";
+    CV_OCL_CODE(CL_INVALID_MIP_LEVEL);
-    case -63: return "CL_INVALID_GLOBAL_WORK_SIZE";
+    CV_OCL_CODE(CL_INVALID_GLOBAL_WORK_SIZE);
-    case -64: return "CL_INVALID_PROPERTY";
+    // OpenCL 1.1
-    case -65: return "CL_INVALID_IMAGE_DESCRIPTOR";
+    CV_OCL_CODE(CL_INVALID_PROPERTY);
-    case -66: return "CL_INVALID_COMPILER_OPTIONS";
+    // OpenCL 1.2
-    case -67: return "CL_INVALID_LINKER_OPTIONS";
+    CV_OCL_CODE(CL_INVALID_IMAGE_DESCRIPTOR);
-    case -68: return "CL_INVALID_DEVICE_PARTITION_COUNT";
+    CV_OCL_CODE(CL_INVALID_COMPILER_OPTIONS);
-    case -69: return "CL_INVALID_PIPE_SIZE";
+    CV_OCL_CODE(CL_INVALID_LINKER_OPTIONS);
-    case -70: return "CL_INVALID_DEVICE_QUEUE";
+    CV_OCL_CODE(CL_INVALID_DEVICE_PARTITION_COUNT);
-    case -1000: return "CL_INVALID_GL_SHAREGROUP_REFERENCE_KHR";
+    // OpenCL 2.0
-    case -1001: return "CL_PLATFORM_NOT_FOUND_KHR";
+    CV_OCL_CODE_(-69, CL_INVALID_PIPE_SIZE);
-    case -1002: return "CL_INVALID_D3D10_DEVICE_KHR";
+    CV_OCL_CODE_(-70, CL_INVALID_DEVICE_QUEUE);
-    case -1003: return "CL_INVALID_D3D10_RESOURCE_KHR";
+    // Extensions
-    case -1004: return "CL_D3D10_RESOURCE_ALREADY_ACQUIRED_KHR";
+    CV_OCL_CODE_(-1000, CL_INVALID_GL_SHAREGROUP_REFERENCE_KHR);
-    case -1005: return "CL_D3D10_RESOURCE_NOT_ACQUIRED_KHR";
+    CV_OCL_CODE_(-1001, CL_PLATFORM_NOT_FOUND_KHR);
-    case -1024: return "clBLAS: Functionality is not implemented";
+    CV_OCL_CODE_(-1002, CL_INVALID_D3D10_DEVICE_KHR);
-    case -1023: return "clBLAS: Library is not initialized yet";
+    CV_OCL_CODE_(-1003, CL_INVALID_D3D10_RESOURCE_KHR);
-    case -1022: return "clBLAS: Matrix A is not a valid memory object";
+    CV_OCL_CODE_(-1004, CL_D3D10_RESOURCE_ALREADY_ACQUIRED_KHR);
-    case -1021: return "clBLAS: Matrix B is not a valid memory object";
+    CV_OCL_CODE_(-1005, CL_D3D10_RESOURCE_NOT_ACQUIRED_KHR);
-    case -1020: return "clBLAS: Matrix C is not a valid memory object";
-    case -1019: return "clBLAS: Vector X is not a valid memory object";
-    case -1018: return "clBLAS: Vector Y is not a valid memory object";
-    case -1017: return "clBLAS: An input dimension (M:N:K) is invalid";
-    case -1016: return "clBLAS: Leading dimension A must not be less than the "
-                       "size of the first dimension";
-    case -1015: return "clBLAS: Leading dimension B must not be less than the "
-                       "size of the second dimension";
-    case -1014: return "clBLAS: Leading dimension C must not be less than the "
-                       "size of the third dimension";
-    case -1013: return "clBLAS: The increment for a vector X must not be 0";
-    case -1012: return "clBLAS: The increment for a vector Y must not be 0";
-    case -1011: return "clBLAS: The memory object for Matrix A is too small";
-    case -1010: return "clBLAS: The memory object for Matrix B is too small";
-    case -1009: return "clBLAS: The memory object for Matrix C is too small";
-    case -1008: return "clBLAS: The memory object for Vector X is too small";
-    case -1007: return "clBLAS: The memory object for Vector Y is too small";
     default: return "Unknown OpenCL error";
     }
+#undef CV_OCL_CODE
+#undef CV_OCL_CODE_
 }
 
 template <typename T>

modules/dnn/CMakeLists.txt

@@ -11,6 +11,14 @@ set(the_description "Deep neural network module. It allows to load models from d
 ocv_add_dispatched_file_force_all("layers/layers_common" AVX AVX2 AVX512_SKX)
 
 ocv_add_module(dnn opencv_core opencv_imgproc WRAP python matlab java js)
+
+ocv_option(OPENCV_DNN_OPENCL "Build with OpenCL support" HAVE_OPENCL)
+if(OPENCV_DNN_OPENCL AND HAVE_OPENCL)
+  add_definitions(-DCV_OCL4DNN=1)
+else()
+  ocv_cmake_hook_append(INIT_MODULE_SOURCES_opencv_dnn "${CMAKE_CURRENT_LIST_DIR}/cmake/hooks/INIT_MODULE_SOURCES_opencv_dnn.cmake")
+endif()
+
 ocv_warnings_disable(CMAKE_CXX_FLAGS -Wno-shadow -Wno-parentheses -Wmaybe-uninitialized -Wsign-promo
                                      -Wmissing-declarations -Wmissing-prototypes
 )
@@ -63,8 +71,15 @@ else()
   set(fw_inc "${CMAKE_CURRENT_LIST_DIR}/misc/caffe" "${CMAKE_CURRENT_LIST_DIR}/misc/tensorflow")
 endif()
 
-ocv_module_include_directories(${fw_inc} ${CMAKE_CURRENT_LIST_DIR}/src/ocl4dnn/include ${OPENCL_INCLUDE_DIRS})
+set(include_dirs ${fw_inc})
-ocv_glob_module_sources(SOURCES ${fw_srcs})
+set(sources_options "")
+if(OPENCV_DNN_OPENCL AND HAVE_OPENCL)
+  list(APPEND include_dirs ${OPENCL_INCLUDE_DIRS})
+else()
+  set(sources_options EXCLUDE_OPENCL)
+endif()
+ocv_module_include_directories(${include_dirs})
+ocv_glob_module_sources(${sources_options} SOURCES ${fw_srcs})
 ocv_create_module(libprotobuf ${LAPACK_LIBRARIES})
 ocv_add_samples()
 ocv_add_accuracy_tests()

modules/dnn/cmake/hooks/INIT_MODULE_SOURCES_opencv_dnn.cmake

@@ -0,0 +1,3 @@
+message(STATUS "opencv_dnn: filter out ocl4dnn source code")
+ocv_list_filterout(OPENCV_MODULE_${the_module}_SOURCES "/ocl4dnn/")
+ocv_list_filterout(OPENCV_MODULE_${the_module}_HEADERS "/ocl4dnn/")

modules/dnn/src/dnn.cpp

@@ -53,6 +53,7 @@
 #include <opencv2/imgproc.hpp>
 
 #include <opencv2/core/utils/configuration.private.hpp>
+#include <opencv2/core/utils/logger.hpp>
 
 namespace cv {
 namespace dnn {
@@ -846,6 +847,13 @@ struct Net::Impl
 
         if (!netWasAllocated || this->blobsToKeep != blobsToKeep_)
         {
+#ifndef HAVE_OPENCL
+            if (preferableBackend == DNN_BACKEND_DEFAULT && preferableTarget == DNN_TARGET_OPENCL)
+            {
+                CV_LOG_WARNING(NULL, "DNN: OpenCL target is not available in this OpenCV build, switching to CPU.")
+                preferableTarget = DNN_TARGET_CPU;
+            }
+#endif
             clear();
 
             allocateLayers(blobsToKeep_);

modules/dnn/src/layers/batch_norm_layer.cpp

@@ -13,7 +13,10 @@ Implementation of Batch Normalization layer.
 #include "op_halide.hpp"
 #include "op_inf_engine.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/concat_layer.cpp

@@ -44,7 +44,10 @@
 #include "layers_common.hpp"
 #include "op_halide.hpp"
 #include "op_inf_engine.hpp"
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/convolution_layer.cpp

@@ -47,9 +47,9 @@
 #include "opencv2/core/hal/hal.hpp"
 #include "opencv2/core/hal/intrin.hpp"
 #include <iostream>
-#include "opencl_kernels_dnn.hpp"
 
 #ifdef HAVE_OPENCL
+#include "opencl_kernels_dnn.hpp"
 using namespace cv::dnn::ocl4dnn;
 #endif
 

modules/dnn/src/layers/detection_output_layer.cpp

@@ -46,7 +46,10 @@
 #include <float.h>
 #include <string>
 #include "../nms.inl.hpp"
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/elementwise_layers.cpp

@@ -46,9 +46,12 @@
 #include "op_inf_engine.hpp"
 #include "opencv2/imgproc.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
-#include "opencl_kernels_dnn.hpp"
 #include <iostream>
 
+#ifdef HAVE_OPENCL
+#include "opencl_kernels_dnn.hpp"
+#endif
+
 namespace cv
 {
 namespace dnn

modules/dnn/src/layers/eltwise_layer.cpp

@@ -44,7 +44,10 @@
 #include "layers_common.hpp"
 #include "op_halide.hpp"
 #include "op_inf_engine.hpp"
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/fully_connected_layer.cpp

@@ -44,10 +44,10 @@
 #include "layers_common.hpp"
 #include "op_halide.hpp"
 #include "op_inf_engine.hpp"
-#include "opencl_kernels_dnn.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
 
 #ifdef HAVE_OPENCL
+#include "opencl_kernels_dnn.hpp"
 using namespace cv::dnn::ocl4dnn;
 #endif
 

modules/dnn/src/layers/layers_common.cpp

@@ -40,6 +40,7 @@
 //
 //M*/
 
+#include "../precomp.hpp"
 #include "layers_common.hpp"
 
 namespace cv

modules/dnn/src/layers/layers_common.hpp

@@ -52,7 +52,7 @@
 #undef CV_CPU_OPTIMIZATION_DECLARATIONS_ONLY
 
 #ifdef HAVE_OPENCL
-#include "ocl4dnn.hpp"
+#include "../ocl4dnn/include/ocl4dnn.hpp"
 #endif
 
 namespace cv

modules/dnn/src/layers/lrn_layer.cpp

@@ -47,10 +47,10 @@
 #include "opencv2/imgproc.hpp"
 #include "opencv2/dnn/shape_utils.hpp"
 #include "opencv2/core/hal/hal.hpp"
-#include "opencl_kernels_dnn.hpp"
 #include <algorithm>
 
 #ifdef HAVE_OPENCL
+#include "opencl_kernels_dnn.hpp"
 using namespace cv::dnn::ocl4dnn;
 #endif
 

modules/dnn/src/layers/mvn_layer.cpp

@@ -43,8 +43,11 @@
 #include "../precomp.hpp"
 #include "layers_common.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
-#include "math_functions.hpp"
+
+#ifdef HAVE_OPENCL
+#include "../ocl4dnn/include/math_functions.hpp"
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/permute_layer.cpp

@@ -45,7 +45,10 @@
 #include "op_inf_engine.hpp"
 #include <float.h>
 #include <algorithm>
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/pooling_layer.cpp

@@ -45,12 +45,13 @@
 #include "opencv2/core/hal/intrin.hpp"
 #include "op_halide.hpp"
 #include "op_inf_engine.hpp"
-#include "opencl_kernels_dnn.hpp"
 #include <float.h>
 #include <algorithm>
 using std::max;
 using std::min;
+
 #ifdef HAVE_OPENCL
+#include "opencl_kernels_dnn.hpp"
 using namespace cv::dnn::ocl4dnn;
 #endif
 
@@ -323,6 +324,8 @@ public:
 
 #if CV_SIMD128
             const int* ofsptr = ofsbuf.empty() ? 0 : (const int*)&ofsbuf[0];
+            if (poolingType == MAX && !compMaxIdx && !ofsptr)
+                CV_Error(Error::StsBadArg, "ofsbuf should be initialized in this mode");
             v_float32x4 idx00(0.f, (float)stride_w, (float)(stride_w*2), (float)(stride_w*3));
             v_float32x4 ones = v_setall_f32(1.f);
             v_float32x4 idx_delta = v_setall_f32((float)(inp_width - kernel_w));

modules/dnn/src/layers/prior_box_layer.cpp

@@ -46,7 +46,10 @@
 #include <float.h>
 #include <algorithm>
 #include <cmath>
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/region_layer.cpp

@@ -44,7 +44,10 @@
 #include <opencv2/dnn/shape_utils.hpp>
 #include <opencv2/dnn/all_layers.hpp>
 #include "nms.inl.hpp"
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/reorg_layer.cpp

@@ -44,7 +44,10 @@
 #include <opencv2/dnn/shape_utils.hpp>
 #include <opencv2/dnn/all_layers.hpp>
 #include <iostream>
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/slice_layer.cpp

@@ -43,7 +43,10 @@
 #include "../precomp.hpp"
 #include "layers_common.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
+
+#ifdef HAVE_OPENCL
 #include "opencl_kernels_dnn.hpp"
+#endif
 
 namespace cv
 {

modules/dnn/src/layers/softmax_layer.cpp

@@ -44,11 +44,12 @@
 #include "layers_common.hpp"
 #include "op_halide.hpp"
 #include "op_inf_engine.hpp"
-#include "opencl_kernels_dnn.hpp"
 #include <algorithm>
 #include <stdlib.h>
 using std::max;
+
 #ifdef HAVE_OPENCL
+#include "opencl_kernels_dnn.hpp"
 using namespace cv::dnn::ocl4dnn;
 #endif
 

modules/dnn/src/ocl4dnn/include/common.hpp

@@ -45,8 +45,6 @@
 #include "../../caffe/glog_emulator.hpp"
 #include <opencv2/core/opencl/runtime/opencl_core.hpp>
 
-#ifdef HAVE_OPENCL
-
 // Macro to select the single (_float) or double (_double) precision kernel
 #define CL_KERNEL_SELECT(kernel) kernel "_float"
 
@@ -58,5 +56,4 @@
 
 bool clOptionSupport(cv::String option);
 
-#endif // HAVE_OPENCL
 #endif

modules/dnn/src/ocl4dnn/include/math_functions.hpp

@@ -52,7 +52,6 @@ namespace dnn
 namespace ocl4dnn
 {
 
-#ifdef HAVE_OPENCL
 enum CBLAS_TRANSPOSE {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113};
 
 template<typename Dtype>
@@ -81,8 +80,6 @@ bool ocl4dnnAXPY(const int32_t N, const Dtype alpha,
                  const UMat x, const int32_t offx, UMat y,
                  const int32_t offy);
 
-#endif  // HAVE_OPENCL
-
 } // namespace ocl4dnn
 } // namespace dnn
 } // namespce cv

modules/dnn/src/ocl4dnn/include/ocl4dnn.hpp

@@ -51,7 +51,6 @@
 #include "common.hpp"
 
 namespace cv { namespace dnn { namespace ocl4dnn {
-#ifdef HAVE_OPENCL
 
 struct OCL4DNNConvConfig
 {
@@ -507,8 +506,7 @@ class OCL4DNNSoftmax
         bool log_softmax_;
         UMat scale_data_;
 };
-#endif // HAVE_OPENCL
+
-} // namespace ocl4dnn
+}}} // namespace cv::dnn::ocl4dnn
-} // namespace dnn
+
-} // namespce cv
 #endif

modules/dnn/src/ocl4dnn/src/common.cpp

@@ -41,17 +41,14 @@
 //M*/
 
 #include "../../precomp.hpp"
-#include "common.hpp"
+#include "../include/common.hpp"
 #include "opencl_kernels_dnn.hpp"
 
 using namespace cv;
 
-#ifdef HAVE_OPENCL
 bool clOptionSupport(cv::String option)
 {
     cv::String errmsg;
     ocl::Program program = ocl::Context::getDefault().getProg(ocl::dnn::dummy_oclsrc, option, errmsg);
     return program.ptr() ? true : false;
 }
-
-#endif // HAVE_OPENCL

modules/dnn/src/ocl4dnn/src/math_functions.cpp

@@ -41,19 +41,13 @@
 //M*/
 
 #include "../../precomp.hpp"
-#include "common.hpp"
+#include "../include/common.hpp"
-#include "math_functions.hpp"
+#include "../include/math_functions.hpp"
 #include <vector>
 #include "opencl_kernels_dnn.hpp"
 
-namespace cv
+namespace cv { namespace dnn { namespace ocl4dnn {
-{
-namespace dnn
-{
-namespace ocl4dnn
-{
 
-#ifdef HAVE_OPENCL
 // Create and copy buffer to image for GEMM's matrix A and B.
 // Will return image to caller if the input image is NULL. Otherwise,
 // will use the image directly. It's caller's responsibility to
@@ -527,8 +521,4 @@ template bool ocl4dnnAXPY<float>(const int32_t N, const float alpha,
                                  const UMat X, const int32_t offX,
                                  UMat Y, const int32_t offY);
 
-#endif  // HAVE_OPENCL
+}}} // namespace cv::dnn::ocl4dnn
-
-} // namespace ocl4dnn
-} // namespace dnn
-} // namespce cv

modules/dnn/src/ocl4dnn/src/ocl4dnn_conv_spatial.cpp

@@ -49,18 +49,17 @@
 #include <fstream>
 #include <sys/stat.h>
 #include <assert.h>
-#include "common.hpp"
+#include "../include/common.hpp"
-#include "ocl4dnn.hpp"
+#include "../include/ocl4dnn.hpp"
 #include "opencl_kernels_dnn.hpp"
-#include "math_functions.hpp"
+#include "../include/math_functions.hpp"
-#include "default_kernel_config.hpp"
+#include "../include/default_kernel_config.hpp"
 
 #if defined WIN32 || defined _WIN32
 #include <windows.h>
 #include <direct.h>
 #endif
 
-#ifdef HAVE_OPENCL
 namespace cv { namespace dnn { namespace ocl4dnn {
 static cv::Mutex kernelConfigMutex;
 typedef std::map<std::string, std::string> kernel_hash_t;
@@ -1855,7 +1854,5 @@ bool OCL4DNNConvSpatial<Dtype>::loadTunedConfig()
 }
 
 template class OCL4DNNConvSpatial<float>;
-} // namespace ocl4dnn
+
-}
+}}} // namespace cv::dnn::ocl4dnn
-}
-#endif // HAVE_OPENCL

modules/dnn/src/ocl4dnn/src/ocl4dnn_inner_product.cpp

@@ -41,11 +41,10 @@
 //M*/
 
 #include "../../precomp.hpp"
-#include "common.hpp"
+#include "../include/common.hpp"
-#include "ocl4dnn.hpp"
+#include "../include/ocl4dnn.hpp"
-#include "math_functions.hpp"
+#include "../include/math_functions.hpp"
 
-#ifdef HAVE_OPENCL
 namespace cv { namespace dnn { namespace ocl4dnn {
 template<typename Dtype>
 OCL4DNNInnerProduct<Dtype>::OCL4DNNInnerProduct(OCL4DNNInnerProductConfig config)
@@ -102,7 +101,5 @@ bool OCL4DNNInnerProduct<Dtype>::Forward(const UMat& bottom,
 }
 
 template class OCL4DNNInnerProduct<float>;
-} // namespace ocl4dnn
+
-}
+}}} // namespace cv::dnn::ocl4dnn
-}
-#endif // HAVE_OPENCL

modules/dnn/src/ocl4dnn/src/ocl4dnn_lrn.cpp

@@ -41,11 +41,10 @@
 //M*/
 
 #include "../../precomp.hpp"
-#include "common.hpp"
+#include "../include/common.hpp"
-#include "ocl4dnn.hpp"
+#include "../include/ocl4dnn.hpp"
 #include "opencl_kernels_dnn.hpp"
 
-#ifdef HAVE_OPENCL
 namespace cv { namespace dnn { namespace ocl4dnn {
 template<typename Dtype>
 OCL4DNNLRN<Dtype>::OCL4DNNLRN(OCL4DNNLRNConfig config)
@@ -119,7 +118,5 @@ bool OCL4DNNLRN<Dtype>::crossChannelForward(const UMat& bottom, UMat& top)
 }
 
 template class OCL4DNNLRN<float>;
-} // namespace ocl4dnn
+
-}
+}}} // namespace cv::dnn::ocl4dnn
-}
-#endif // HAVE_OPENCL

modules/dnn/src/ocl4dnn/src/ocl4dnn_pool.cpp

@@ -42,11 +42,10 @@
 #include "../../precomp.hpp"
 #include <string>
 #include <vector>
-#include "common.hpp"
+#include "../include/common.hpp"
-#include "ocl4dnn.hpp"
+#include "../include/ocl4dnn.hpp"
 #include "opencl_kernels_dnn.hpp"
 
-#ifdef HAVE_OPENCL
 namespace cv { namespace dnn { namespace ocl4dnn {
 template<typename Dtype>
 OCL4DNNPool<Dtype>::OCL4DNNPool(OCL4DNNPoolConfig config)
@@ -208,7 +207,5 @@ bool OCL4DNNPool<Dtype>::Forward(const UMat& bottom,
 }
 
 template class OCL4DNNPool<float>;
-} // namespace ocl4dnn
+
-}
+}}} // namespace cv::dnn::ocl4dnn
-}
-#endif // HAVE_OPENCL

modules/dnn/src/ocl4dnn/src/ocl4dnn_softmax.cpp

@@ -41,11 +41,10 @@
 
 #include "../../precomp.hpp"
 #include <vector>
-#include "common.hpp"
+#include "../include/common.hpp"
-#include "ocl4dnn.hpp"
+#include "../include/ocl4dnn.hpp"
 #include "opencl_kernels_dnn.hpp"
 
-#ifdef HAVE_OPENCL
 namespace cv { namespace dnn { namespace ocl4dnn {
 template<typename Dtype>
 OCL4DNNSoftmax<Dtype>::OCL4DNNSoftmax(OCL4DNNSoftmaxConfig config)
@@ -130,7 +129,5 @@ bool OCL4DNNSoftmax<Dtype>::Forward(const UMat& bottom, UMat& top)
 }
 
 template class OCL4DNNSoftmax<float>;
-} // namespace ocl4dnn
+
-}
+}}} // namespace cv::dnn::ocl4dnn
-}
-#endif // HAVE_OPENCL

modules/dnn/src/precomp.hpp

@@ -40,13 +40,28 @@
 //M*/
 
 #include <opencv2/core.hpp>
+#include "cvconfig.h"
+
+#ifndef CV_OCL4DNN
+#define CV_OCL4DNN 0
+#endif
+
+#if CV_OCL4DNN
+#ifndef HAVE_OPENCL
+#error "Configuration error: re-run CMake from clean build directory"
+#endif
+#else
+#undef HAVE_OPENCL
+#endif
 #include <opencv2/core/ocl.hpp>
 #include <opencv2/core/opencl/ocl_defs.hpp>
+
+
 #include <opencv2/core/utils/trace.hpp>
-#include "cvconfig.h"
 #include <opencv2/dnn.hpp>
 #include <opencv2/dnn/all_layers.hpp>
 
+
 namespace cv { namespace dnn {
 CV__DNN_EXPERIMENTAL_NS_BEGIN
 Mutex& getInitializationMutex();

modules/imgcodecs/src/grfmt_tiff.cpp

@@ -775,6 +775,7 @@ bool TiffEncoder::writeLibTiff( const std::vector<Mat>& img_vec, const std::vect
             }
             default:
             {
+                TIFFClose(pTiffHandle);
                 return false;
             }
         }

modules/imgproc/src/connectedcomponents.cpp

@@ -579,9 +579,6 @@ namespace cv{
             CV_Assert(img.cols == imgLabels.cols);
             CV_Assert(connectivity == 8 || connectivity == 4);
 
-            const int nThreads = cv::getNumberOfCPUs();
-            cv::setNumThreads(nThreads);
-
             const int h = img.rows;
             const int w = img.cols;
 
@@ -606,12 +603,13 @@ namespace cv{
             P[0] = 0;
 
             cv::Range range(0, h);
+            const double nParallelStripes = std::max(1, std::min(h / 2, getNumThreads()*4));
+
             LabelT nLabels = 1;
 
             if (connectivity == 8){
-                //First scan, each thread works with chunk of img.rows/nThreads rows
+                //First scan
-                //e.g. 300 rows, 4 threads -> each chunks is composed of 75 rows
+                cv::parallel_for_(range, FirstScan8Connectivity(img, imgLabels, P, chunksSizeAndLabels), nParallelStripes);
-                cv::parallel_for_(range, FirstScan8Connectivity(img, imgLabels, P, chunksSizeAndLabels), nThreads);
 
                 //merge labels of different chunks
                 mergeLabels8Connectivity(imgLabels, P, chunksSizeAndLabels);
@@ -621,9 +619,8 @@ namespace cv{
                 }
             }
             else{
-                //First scan, each thread works with chunk of img.rows/nThreads rows
+                //First scan
-                //e.g. 300 rows, 4 threads -> each chunks is composed of 75 rows
+                cv::parallel_for_(range, FirstScan4Connectivity(img, imgLabels, P, chunksSizeAndLabels), nParallelStripes);
-                cv::parallel_for_(range, FirstScan4Connectivity(img, imgLabels, P, chunksSizeAndLabels), nThreads);
 
                 //merge labels of different chunks
                 mergeLabels4Connectivity(imgLabels, P, chunksSizeAndLabels);
@@ -638,7 +635,7 @@ namespace cv{
 
             sop.init(nLabels);
             //Second scan
-            cv::parallel_for_(range, SecondScan(imgLabels, P, sop, sopArray, nLabels), nThreads);
+            cv::parallel_for_(range, SecondScan(imgLabels, P, sop, sopArray, nLabels), nParallelStripes);
             StatsOp::mergeStats(imgLabels, sopArray, sop, nLabels);
             sop.finish();
 
@@ -2530,9 +2527,6 @@ namespace cv{
             CV_Assert(img.cols == imgLabels.cols);
             CV_Assert(connectivity == 8);
 
-            const int nThreads = cv::getNumberOfCPUs();
-            cv::setNumThreads(nThreads);
-
             const int h = img.rows;
             const int w = img.cols;
 
@@ -2556,10 +2550,11 @@ namespace cv{
             P[0] = 0;
 
             cv::Range range(0, h);
+            const double nParallelStripes = std::max(1, std::min(h / 2, getNumThreads()*4));
 
             //First scan, each thread works with chunk of img.rows/nThreads rows
             //e.g. 300 rows, 4 threads -> each chunks is composed of 75 rows
-            cv::parallel_for_(range, FirstScan(img, imgLabels, P, chunksSizeAndLabels), nThreads);
+            cv::parallel_for_(range, FirstScan(img, imgLabels, P, chunksSizeAndLabels), nParallelStripes);
 
             //merge labels of different chunks
             mergeLabels(img, imgLabels, P, chunksSizeAndLabels);
@@ -2574,7 +2569,7 @@ namespace cv{
             sop.init(nLabels);
 
             //Second scan
-            cv::parallel_for_(range, SecondScan(img, imgLabels, P, sop, sopArray, nLabels), nThreads);
+            cv::parallel_for_(range, SecondScan(img, imgLabels, P, sop, sopArray, nLabels), nParallelStripes);
 
             StatsOp::mergeStats(imgLabels, sopArray, sop, nLabels);
             sop.finish();
@@ -3936,12 +3931,12 @@ namespace cv{
         int lDepth = L.depth();
         int iDepth = I.depth();
         const char *currentParallelFramework = cv::currentParallelFramework();
-        const int numberOfCPUs = cv::getNumberOfCPUs();
+        const int nThreads = cv::getNumThreads();
 
         CV_Assert(iDepth == CV_8U || iDepth == CV_8S);
 
-        //Run parallel labeling only if the rows of the image are at least twice the number returned by getNumberOfCPUs
+        //Run parallel labeling only if the rows of the image are at least twice the number of available threads
-        const bool is_parallel = currentParallelFramework != NULL && numberOfCPUs > 1 && L.rows / numberOfCPUs >= 2;
+        const bool is_parallel = currentParallelFramework != NULL && nThreads > 1 && L.rows / nThreads >= 2;
 
         if (ccltype == CCL_WU || connectivity == 4){
             // Wu algorithm is used

samples/cpp/bgfg_segm.cpp

@@ -1,3 +1,7 @@
+// This file is part of OpenCV project.
+// It is subject to the license terms in the LICENSE file found in the top-level directory
+// of this distribution and at http://opencv.org/license.html
+
 #include "opencv2/core.hpp"
 #include "opencv2/imgproc.hpp"
 #include "opencv2/video.hpp"
@@ -10,10 +14,10 @@ using namespace cv;
 
 int main(int argc, const char** argv)
 {
-    const String keys = "{c camera||use video stream from camera (default is NO)}"
+    const String keys = "{c camera     | 0 | use video stream from camera (device index starting from 0) }"
-                        "{fn file_name|../data/tree.avi|video file}"
+                        "{fn file_name |   | use video file as input }"
-                        "{m method|mog2|method: background subtraction algorithm ('knn', 'mog2')}"
+                        "{m method | mog2 | method: background subtraction algorithm ('knn', 'mog2')}"
-                        "{h help||show help message}";
+                        "{h help | | show help message}";
     CommandLineParser parser(argc, argv, keys);
     parser.about("This sample demonstrates background segmentation.");
     if (parser.has("help"))
@@ -21,7 +25,7 @@ int main(int argc, const char** argv)
         parser.printMessage();
         return 0;
     }
-    bool useCamera = parser.has("camera");
+    int camera = parser.get<int>("camera");
     String file = parser.get<String>("file_name");
     String method = parser.get<String>("method");
     if (!parser.check())
@@ -31,13 +35,13 @@ int main(int argc, const char** argv)
     }
 
     VideoCapture cap;
-    if (useCamera)
+    if (file.empty())
-        cap.open(0);
+        cap.open(camera);
     else
         cap.open(file.c_str());
     if (!cap.isOpened())
     {
-        cout << "Can not open video stream: '" << (useCamera ? "<camera 0>" : file) << "'" << endl;
+        cout << "Can not open video stream: '" << (file.empty() ? "<camera>" : file) << "'" << endl;
         return 2;
     }
 

samples/cpp/opencv_version.cpp

@@ -1,16 +1,17 @@
+// This file is part of OpenCV project.
+// It is subject to the license terms in the LICENSE file found in the top-level directory
+// of this distribution and at http://opencv.org/license.html
+
 #include <opencv2/core/utility.hpp>
 #include <iostream>
 
-const char* keys =
+static const std::string keys = "{ b build | | print complete build info }"
-{
+                                "{ h help  | | print this help           }";
-    "{ b build | | print complete build info }"
-    "{ h help  | | print this help           }"
-};
 
 int main(int argc, const char* argv[])
 {
     cv::CommandLineParser parser(argc, argv, keys);
-
+    parser.about("This sample outputs OpenCV version and build configuration.");
     if (parser.has("help"))
     {
         parser.printMessage();
@@ -27,6 +28,5 @@ int main(int argc, const char* argv[])
     {
         std::cout << "Welcome to OpenCV " << CV_VERSION << std::endl;
     }
-
     return 0;
 }

samples/cpp/peopledetect.cpp

@@ -1,177 +1,126 @@
-#include <iostream>
+// This file is part of OpenCV project.
-#include <stdexcept>
+// It is subject to the license terms in the LICENSE file found in the top-level directory
+// of this distribution and at http://opencv.org/license.html
+
 #include <opencv2/objdetect.hpp>
 #include <opencv2/highgui.hpp>
 #include <opencv2/imgproc.hpp>
-#include <opencv2/imgcodecs.hpp>
-#include <opencv2/video.hpp>
 #include <opencv2/videoio.hpp>
+#include <iostream>
+#include <iomanip>
 
 using namespace cv;
 using namespace std;
 
-
+class Detector
-const char* keys =
 {
-    "{ help h      |                     | print help message }"
+    enum Mode { Default, Daimler } m;
-    "{ image i     |                     | specify input image}"
+    HOGDescriptor hog, hog_d;
-    "{ camera c    |                     | enable camera capturing }"
+public:
-    "{ video v     | ../data/vtest.avi   | use video as input }"
+    Detector() : m(Default), hog(), hog_d(Size(48, 96), Size(16, 16), Size(8, 8), Size(8, 8), 9)
-    "{ directory d |                     | images directory}"
-};
-
-static void detectAndDraw(const HOGDescriptor &hog, Mat &img)
-{
-    vector<Rect> found, found_filtered;
-    double t = (double) getTickCount();
-    // Run the detector with default parameters. to get a higher hit-rate
-    // (and more false alarms, respectively), decrease the hitThreshold and
-    // groupThreshold (set groupThreshold to 0 to turn off the grouping completely).
-    hog.detectMultiScale(img, found, 0, Size(8,8), Size(32,32), 1.05, 2);
-    t = (double) getTickCount() - t;
-    cout << "detection time = " << (t*1000./cv::getTickFrequency()) << " ms" << endl;
-
-    for(size_t i = 0; i < found.size(); i++ )
     {
-        Rect r = found[i];
+        hog.setSVMDetector(HOGDescriptor::getDefaultPeopleDetector());
-
+        hog_d.setSVMDetector(HOGDescriptor::getDaimlerPeopleDetector());
-        size_t j;
-        // Do not add small detections inside a bigger detection.
-        for ( j = 0; j < found.size(); j++ )
-            if ( j != i && (r & found[j]) == r )
-                break;
-
-        if ( j == found.size() )
-            found_filtered.push_back(r);
     }
-
+    void toggleMode() { m = (m == Default ? Daimler : Default); }
-    for (size_t i = 0; i < found_filtered.size(); i++)
+    string modeName() const { return (m == Default ? "Default" : "Daimler"); }
+    vector<Rect> detect(InputArray img)
+    {
+        // Run the detector with default parameters. to get a higher hit-rate
+        // (and more false alarms, respectively), decrease the hitThreshold and
+        // groupThreshold (set groupThreshold to 0 to turn off the grouping completely).
+        vector<Rect> found;
+        if (m == Default)
+            hog.detectMultiScale(img, found, 0, Size(8,8), Size(32,32), 1.05, 2, false);
+        else if (m == Daimler)
+            hog_d.detectMultiScale(img, found, 0.5, Size(8,8), Size(32,32), 1.05, 2, true);
+        return found;
+    }
+    void adjustRect(Rect & r) const
     {
-        Rect r = found_filtered[i];
-
         // The HOG detector returns slightly larger rectangles than the real objects,
         // so we slightly shrink the rectangles to get a nicer output.
         r.x += cvRound(r.width*0.1);
         r.width = cvRound(r.width*0.8);
         r.y += cvRound(r.height*0.07);
         r.height = cvRound(r.height*0.8);
-        rectangle(img, r.tl(), r.br(), cv::Scalar(0,255,0), 3);
     }
-}
+};
+
+static const string keys = "{ help h   |   | print help message }"
+                           "{ camera c | 0 | capture video from camera (device index starting from 0) }"
+                           "{ video v  |   | use video as input }";
 
 int main(int argc, char** argv)
 {
     CommandLineParser parser(argc, argv, keys);
-
+    parser.about("This sample demonstrates the use ot the HoG descriptor.");
     if (parser.has("help"))
     {
-        cout << "\nThis program demonstrates the use of the HoG descriptor using\n"
-            " HOGDescriptor::hog.setSVMDetector(HOGDescriptor::getDefaultPeopleDetector());\n";
         parser.printMessage();
-        cout << "During execution:\n\tHit q or ESC key to quit.\n"
-            "\tUsing OpenCV version " << CV_VERSION << "\n"
-            "Note: camera device number must be different from -1.\n" << endl;
         return 0;
     }
-
+    int camera = parser.get<int>("camera");
-    HOGDescriptor hog;
+    string file = parser.get<string>("video");
-    hog.setSVMDetector(HOGDescriptor::getDefaultPeopleDetector());
+    if (!parser.check())
-    namedWindow("people detector", 1);
-
-    string pattern_glob = "";
-    string video_filename = "../data/vtest.avi";
-    int camera_id = -1;
-    if (parser.has("directory"))
     {
-        pattern_glob = parser.get<string>("directory");
+        parser.printErrors();
-    }
+        return 1;
-    else if (parser.has("image"))
-    {
-        pattern_glob = parser.get<string>("image");
-    }
-    else if (parser.has("camera"))
-    {
-        camera_id = parser.get<int>("camera");
-    }
-    else if (parser.has("video"))
-    {
-        video_filename = parser.get<string>("video");
     }
 
-    if (!pattern_glob.empty() || camera_id != -1 || !video_filename.empty())
+    VideoCapture cap;
+    if (file.empty())
+        cap.open(camera);
+    else
+        cap.open(file.c_str());
+    if (!cap.isOpened())
     {
-        //Read from input image files
+        cout << "Can not open video stream: '" << (file.empty() ? "<camera>" : file) << "'" << endl;
-        vector<String> filenames;
+        return 2;
-        //Read from video file
+    }
-        VideoCapture vc;
-        Mat frame;
 
-        if (!pattern_glob.empty())
+    cout << "Press 'q' or <ESC> to quit." << endl;
+    cout << "Press <space> to toggle between Default and Daimler detector" << endl;
+    Detector detector;
+    Mat frame;
+    for (;;)
+    {
+        cap >> frame;
+        if (frame.empty())
         {
-            String folder(pattern_glob);
+            cout << "Finished reading: empty frame" << endl;
-            glob(folder, filenames);
+            break;
         }
-        else if (camera_id != -1)
+        int64 t = getTickCount();
+        vector<Rect> found = detector.detect(frame);
+        t = getTickCount() - t;
+
+        // show the window
         {
-            vc.open(camera_id);
+            ostringstream buf;
-            if (!vc.isOpened())
+            buf << "Mode: " << detector.modeName() << " ||| "
-            {
+                << "FPS: " << fixed << setprecision(1) << (getTickFrequency() / (double)t);
-                stringstream msg;
+            putText(frame, buf.str(), Point(10, 30), FONT_HERSHEY_PLAIN, 2.0, Scalar(0, 0, 255), 2, LINE_AA);
-                msg << "can't open camera: " << camera_id;
-                throw runtime_error(msg.str());
-            }
         }
-        else
+        for (vector<Rect>::iterator i = found.begin(); i != found.end(); ++i)
         {
-            vc.open(video_filename.c_str());
+            Rect &r = *i;
-            if (!vc.isOpened())
+            detector.adjustRect(r);
-                throw runtime_error(string("can't open video file: " + video_filename));
+            rectangle(frame, r.tl(), r.br(), cv::Scalar(0, 255, 0), 2);
         }
+        imshow("People detector", frame);
 
-        vector<String>::const_iterator it_image = filenames.begin();
+        // interact with user
-
+        const char key = (char)waitKey(30);
-        for (;;)
+        if (key == 27 || key == 'q') // ESC
         {
-            if (!pattern_glob.empty())
+            cout << "Exit requested" << endl;
-            {
+            break;
-                bool read_image_ok = false;
+        }
-                for (; it_image != filenames.end(); ++it_image)
+        else if (key == ' ')
-                {
+        {
-                    cout << "\nRead: " << *it_image << endl;
+            detector.toggleMode();
-                    // Read current image
-                    frame = imread(*it_image);
-
-                    if (!frame.empty())
-                    {
-                        ++it_image;
-                        read_image_ok = true;
-                        break;
-                    }
-                }
-
-                //No more valid images
-                if (!read_image_ok)
-                {
-                    //Release the image in order to exit the while loop
-                    frame.release();
-                }
-            }
-            else
-            {
-                vc >> frame;
-            }
-
-            if (frame.empty())
-                break;
-
-            detectAndDraw(hog, frame);
-
-            imshow("people detector", frame);
-            int c = waitKey( vc.isOpened() ? 30 : 0 ) & 255;
-            if ( c == 'q' || c == 'Q' || c == 27)
-                break;
         }
     }
-
     return 0;
 }

samples/dnn/colorization.cpp

@@ -1,20 +1,18 @@
-//
+// This file is part of OpenCV project.
-// This program is based on https://github.com/richzhang/colorization/blob/master/colorization/colorize.py
+// It is subject to the license terms in the LICENSE file found in the top-level directory
-//  download the caffemodel from: http://eecs.berkeley.edu/~rich.zhang/projects/2016_colorization/files/demo_v2/colorization_release_v2.caffemodel
+// of this distribution and at http://opencv.org/license.html
-//  and the prototxt from: https://github.com/richzhang/colorization/blob/master/colorization/models/colorization_deploy_v2.prototxt
+
-//
 #include <opencv2/dnn.hpp>
 #include <opencv2/imgproc.hpp>
 #include <opencv2/highgui.hpp>
+#include <iostream>
+
 using namespace cv;
 using namespace cv::dnn;
-
-#include <iostream>
 using namespace std;
 
-
 // the 313 ab cluster centers from pts_in_hull.npy (already transposed)
-float hull_pts[] = {
+static float hull_pts[] = {
     -90., -90., -90., -90., -90., -80., -80., -80., -80., -80., -80., -80., -80., -70., -70., -70., -70., -70., -70., -70., -70.,
     -70., -70., -60., -60., -60., -60., -60., -60., -60., -60., -60., -60., -60., -60., -50., -50., -50., -50., -50., -50., -50., -50.,
     -50., -50., -50., -50., -50., -50., -40., -40., -40., -40., -40., -40., -40., -40., -40., -40., -40., -40., -40., -40., -40., -30.,
@@ -43,54 +41,61 @@ float hull_pts[] = {
     -20., -10., 0., 10., 20., 30., 40., 50., 60., 70., -90., -80., -70., -60., -50., -40., -30., -20., -10., 0.
 };
 
-
 int main(int argc, char **argv)
 {
-    CommandLineParser parser(argc, argv,
+    const string about =
-        "{ help           | false | print this help message }"
+        "This sample demonstrates recoloring grayscale images with dnn.\n"
-        "{ proto          | colorization_deploy_v2.prototxt | model configuration }"
+        "This program is based on:\n"
-        "{ model          | colorization_release_v2.caffemodel | model weights }"
+        "  http://richzhang.github.io/colorization\n"
-        "{ image          | space_shuttle.jpg | path to image file }"
+        "  https://github.com/richzhang/colorization\n"
-        "{ opencl         | false | enable OpenCL }"
+        "Download caffemodel and prototxt files:\n"
-    );
+        "  http://eecs.berkeley.edu/~rich.zhang/projects/2016_colorization/files/demo_v2/colorization_release_v2.caffemodel\n"
-
+        "  https://raw.githubusercontent.com/richzhang/colorization/master/colorization/models/colorization_deploy_v2.prototxt\n";
-    String modelTxt = parser.get<string>("proto");
+    const string keys =
-    String modelBin = parser.get<string>("model");
+        "{ h help |                                    | print this help message }"
-    String imageFile = parser.get<String>("image");
+        "{ proto  | colorization_deploy_v2.prototxt    | model configuration }"
-    if (parser.get<bool>("help") || modelTxt.empty() || modelBin.empty() || imageFile.empty())
+        "{ model  | colorization_release_v2.caffemodel | model weights }"
+        "{ image  | space_shuttle.jpg                  | path to image file }"
+        "{ opencl |                                    | enable OpenCL }";
+    CommandLineParser parser(argc, argv, keys);
+    parser.about(about);
+    if (parser.has("help"))
     {
-        cout << "A sample app to demonstrate recoloring grayscale images with dnn." << endl;
         parser.printMessage();
         return 0;
     }
-
+    string modelTxt = parser.get<string>("proto");
-    // fixed input size for the pretrained network
+    string modelBin = parser.get<string>("model");
-    int W_in = 224;
+    string imageFile = parser.get<string>("image");
-    int H_in = 224;
+    bool useOpenCL = parser.has("opencl");
-
+    if (!parser.check())
-    Net net = dnn::readNetFromCaffe(modelTxt, modelBin);
-
-    // setup additional layers:
-    int sz[] = {2, 313, 1, 1};
-    Mat pts_in_hull(4, sz, CV_32F, hull_pts);
-    Ptr<dnn::Layer> class8_ab = net.getLayer("class8_ab");
-    class8_ab->blobs.push_back(pts_in_hull);
-
-    Ptr<dnn::Layer> conv8_313_rh = net.getLayer("conv8_313_rh");
-    conv8_313_rh->blobs.push_back(Mat(1, 313, CV_32F, 2.606f));
-
-    if (parser.get<bool>("opencl"))
     {
-        net.setPreferableTarget(DNN_TARGET_OPENCL);
+        parser.printErrors();
+        return 1;
     }
 
     Mat img = imread(imageFile);
     if (img.empty())
     {
-        std::cerr << "Can't read image from the file: " << imageFile << std::endl;
+        cout << "Can't read image from file: " << imageFile << endl;
-        exit(-1);
+        return 2;
     }
 
+    // fixed input size for the pretrained network
+    const int W_in = 224;
+    const int H_in = 224;
+    Net net = dnn::readNetFromCaffe(modelTxt, modelBin);
+    if (useOpenCL)
+        net.setPreferableTarget(DNN_TARGET_OPENCL);
+
+    // setup additional layers:
+    int sz[] = {2, 313, 1, 1};
+    const Mat pts_in_hull(4, sz, CV_32F, hull_pts);
+    Ptr<dnn::Layer> class8_ab = net.getLayer("class8_ab");
+    class8_ab->blobs.push_back(pts_in_hull);
+    Ptr<dnn::Layer> conv8_313_rh = net.getLayer("conv8_313_rh");
+    conv8_313_rh->blobs.push_back(Mat(1, 313, CV_32F, Scalar(2.606)));
+
     // extract L channel and subtract mean
     Mat lab, L, input;
     img.convertTo(img, CV_32F, 1.0/255);
@@ -111,13 +116,11 @@ int main(int argc, char **argv)
     resize(a, a, img.size());
     resize(b, b, img.size());
 
-    // merge, and convert back to bgr
+    // merge, and convert back to BGR
     Mat color, chn[] = {L, a, b};
     merge(chn, 3, lab);
     cvtColor(lab, color, COLOR_Lab2BGR);
 
-    namedWindow("color", WINDOW_NORMAL);
-    namedWindow("original", WINDOW_NORMAL);
     imshow("color", color);
     imshow("original", img);
     waitKey();

samples/tapi/dense_optical_flow.cpp

@@ -0,0 +1,151 @@
+// This file is part of OpenCV project.
+// It is subject to the license terms in the LICENSE file found in the top-level directory
+// of this distribution and at http://opencv.org/license.html
+
+#include <iostream>
+#include <iomanip>
+#include <vector>
+
+#include "opencv2/core/ocl.hpp"
+#include "opencv2/core/utility.hpp"
+#include "opencv2/imgcodecs.hpp"
+#include "opencv2/videoio.hpp"
+#include "opencv2/highgui.hpp"
+#include "opencv2/video.hpp"
+
+using namespace std;
+using namespace cv;
+
+static Mat getVisibleFlow(InputArray flow)
+{
+    vector<UMat> flow_vec;
+    split(flow, flow_vec);
+    UMat magnitude, angle;
+    cartToPolar(flow_vec[0], flow_vec[1], magnitude, angle, true);
+    magnitude.convertTo(magnitude, CV_32F, 0.2);
+    vector<UMat> hsv_vec;
+    hsv_vec.push_back(angle);
+    hsv_vec.push_back(UMat::ones(angle.size(), angle.type()));
+    hsv_vec.push_back(magnitude);
+    UMat hsv;
+    merge(hsv_vec, hsv);
+    Mat img;
+    cvtColor(hsv, img, COLOR_HSV2BGR);
+    return img;
+}
+
+static Size fitSize(const Size & sz,  const Size & bounds)
+{
+    CV_Assert(sz.area() > 0);
+    if (sz.width > bounds.width || sz.height > bounds.height)
+    {
+        double scale = std::min((double)bounds.width / sz.width, (double)bounds.height / sz.height);
+        return Size(cvRound(sz.width * scale), cvRound(sz.height * scale));
+    }
+    return sz;
+}
+
+int main(int argc, const char* argv[])
+{
+    const char* keys =
+            "{ h help     |     | print help message }"
+            "{ c camera   | 0   | capture video from camera (device index starting from 0) }"
+            "{ a algorithm | fb | algorithm (supported: 'fb', 'tvl')}"
+            "{ m cpu      |     | run without OpenCL }"
+            "{ v video    |     | use video as input }"
+            "{ o original |     | use original frame size (do not resize to 640x480)}"
+            ;
+    CommandLineParser parser(argc, argv, keys);
+    parser.about("This sample demonstrates using of dense optical flow algorithms.");
+    if (parser.has("help"))
+    {
+        parser.printMessage();
+        return 0;
+    }
+    int camera = parser.get<int>("camera");
+    string algorithm = parser.get<string>("algorithm");
+    bool useCPU = parser.has("cpu");
+    string filename = parser.get<string>("video");
+    bool useOriginalSize = parser.has("original");
+    if (!parser.check())
+    {
+        parser.printErrors();
+        return 1;
+    }
+
+    VideoCapture cap;
+    if(filename.empty())
+        cap.open(camera);
+    else
+        cap.open(filename);
+    if (!cap.isOpened())
+    {
+        cout << "Can not open video stream: '" << (filename.empty() ? "<camera>" : filename) << "'" << endl;
+        return 2;
+    }
+
+    cv::Ptr<cv::DenseOpticalFlow> alg;
+    if (algorithm == "fb")
+        alg = cv::FarnebackOpticalFlow::create();
+    else if (algorithm == "tvl")
+        alg = cv::DualTVL1OpticalFlow::create();
+    else
+    {
+        cout << "Invalid algorithm: " << algorithm << endl;
+        return 3;
+    }
+
+    ocl::setUseOpenCL(!useCPU);
+
+    cout << "Press 'm' to toggle CPU/GPU processing mode" << endl;
+    cout << "Press ESC or 'q' to exit" << endl;
+
+    UMat prevFrame, frame, input_frame, flow;
+    for(;;)
+    {
+        if (!cap.read(input_frame) || input_frame.empty())
+        {
+            cout << "Finished reading: empty frame" << endl;
+            break;
+        }
+        Size small_size = fitSize(input_frame.size(), Size(640, 480));
+        if (!useOriginalSize && small_size != input_frame.size())
+            resize(input_frame, frame, small_size);
+        else
+            frame = input_frame;
+        cvtColor(frame, frame, COLOR_BGR2GRAY);
+        imshow("frame", frame);
+        if (!prevFrame.empty())
+        {
+            int64 t = getTickCount();
+            alg->calc(prevFrame, frame, flow);
+            t = getTickCount() - t;
+            {
+                Mat img = getVisibleFlow(flow);
+                ostringstream buf;
+                buf << "Algo: " << algorithm << " | "
+                    << "Mode: " << (useCPU ? "CPU" : "GPU") << " | "
+                    << "FPS: " << fixed << setprecision(1) << (getTickFrequency() / (double)t);
+                putText(img, buf.str(), Point(10, 30), FONT_HERSHEY_PLAIN, 2.0, Scalar(0, 0, 255), 2, LINE_AA);
+                imshow("Dense optical flow field", img);
+            }
+        }
+        frame.copyTo(prevFrame);
+
+        // interact with user
+        const char key = (char)waitKey(30);
+        if (key == 27 || key == 'q') // ESC
+        {
+            cout << "Exit requested" << endl;
+            break;
+        }
+        else if (key == 'm')
+        {
+            useCPU = !useCPU;
+            ocl::setUseOpenCL(!useCPU);
+            cout << "Set processing mode to: " << (useCPU ? "CPU" : "GPU") << endl;
+        }
+    }
+
+    return 0;
+}

samples/tapi/opencl_custom_kernel.cpp

@@ -1,3 +1,7 @@
+// This file is part of OpenCV project.
+// It is subject to the license terms in the LICENSE file found in the top-level directory
+// of this distribution and at http://opencv.org/license.html
+
 #include "opencv2/core.hpp"
 #include "opencv2/core/ocl.hpp"
 #include "opencv2/highgui.hpp"

samples/tapi/tvl1_optical_flow.cpp

@@ -1,233 +0,0 @@
-#include <iostream>
-#include <vector>
-#include <iomanip>
-
-#include "opencv2/core/ocl.hpp"
-#include "opencv2/core/utility.hpp"
-#include "opencv2/imgcodecs.hpp"
-#include "opencv2/videoio.hpp"
-#include "opencv2/highgui.hpp"
-#include "opencv2/video.hpp"
-
-using namespace std;
-using namespace cv;
-
-typedef unsigned char uchar;
-#define LOOP_NUM 10
-int64 work_begin = 0;
-int64 work_end = 0;
-
-static void workBegin()
-{
-    work_begin = getTickCount();
-}
-static void workEnd()
-{
-    work_end += (getTickCount() - work_begin);
-}
-static double getTime()
-{
-    return work_end * 1000. / getTickFrequency();
-}
-
-template <typename T> inline T clamp (T x, T a, T b)
-{
-    return ((x) > (a) ? ((x) < (b) ? (x) : (b)) : (a));
-}
-
-template <typename T> inline T mapValue(T x, T a, T b, T c, T d)
-{
-    x = ::clamp(x, a, b);
-    return c + (d - c) * (x - a) / (b - a);
-}
-
-static void getFlowField(const Mat& u, const Mat& v, Mat& flowField)
-{
-    float maxDisplacement = 1.0f;
-
-    for (int i = 0; i < u.rows; ++i)
-    {
-        const float* ptr_u = u.ptr<float>(i);
-        const float* ptr_v = v.ptr<float>(i);
-
-        for (int j = 0; j < u.cols; ++j)
-        {
-            float d = max(fabsf(ptr_u[j]), fabsf(ptr_v[j]));
-
-            if (d > maxDisplacement)
-                maxDisplacement = d;
-        }
-    }
-
-    flowField.create(u.size(), CV_8UC4);
-
-    for (int i = 0; i < flowField.rows; ++i)
-    {
-        const float* ptr_u = u.ptr<float>(i);
-        const float* ptr_v = v.ptr<float>(i);
-
-
-        Vec4b* row = flowField.ptr<Vec4b>(i);
-
-        for (int j = 0; j < flowField.cols; ++j)
-        {
-            row[j][0] = 0;
-            row[j][1] = static_cast<unsigned char> (mapValue (-ptr_v[j], -maxDisplacement, maxDisplacement, 0.0f, 255.0f));
-            row[j][2] = static_cast<unsigned char> (mapValue ( ptr_u[j], -maxDisplacement, maxDisplacement, 0.0f, 255.0f));
-            row[j][3] = 255;
-        }
-    }
-}
-
-
-int main(int argc, const char* argv[])
-{
-    const char* keys =
-        "{ h help     |                 | print help message }"
-        "{ l left     |                 | specify left image }"
-        "{ r right    |                 | specify right image }"
-        "{ o output   | tvl1_output.jpg | specify output save path }"
-        "{ c camera   | 0               | enable camera capturing }"
-        "{ m cpu_mode |                 | run without OpenCL }"
-        "{ v video    |                 | use video as input }";
-
-    CommandLineParser cmd(argc, argv, keys);
-
-    if (cmd.has("help"))
-    {
-        cout << "Usage: pyrlk_optical_flow [options]" << endl;
-        cout << "Available options:" << endl;
-        cmd.printMessage();
-        return EXIT_SUCCESS;
-    }
-
-    string fname0 = cmd.get<string>("l");
-    string fname1 = cmd.get<string>("r");
-    string vdofile = cmd.get<string>("v");
-    string outpath = cmd.get<string>("o");
-    bool useCPU = cmd.get<bool>("m");
-    bool useCamera = cmd.get<bool>("c");
-    int inputName = cmd.get<int>("c");
-
-    UMat frame0, frame1;
-    imread(fname0, cv::IMREAD_GRAYSCALE).copyTo(frame0);
-    imread(fname1, cv::IMREAD_GRAYSCALE).copyTo(frame1);
-    cv::Ptr<cv::DenseOpticalFlow> alg = cv::createOptFlow_DualTVL1();
-
-    UMat flow;
-    Mat show_flow;
-    vector<UMat> flow_vec;
-    if (frame0.empty() || frame1.empty())
-        useCamera = true;
-
-    if (useCamera)
-    {
-        VideoCapture capture;
-        UMat frame, frameCopy;
-        UMat frame0Gray, frame1Gray;
-        UMat ptr0, ptr1;
-
-        if(vdofile.empty())
-            capture.open( inputName );
-        else
-            capture.open(vdofile.c_str());
-
-        if(!capture.isOpened())
-        {
-            if(vdofile.empty())
-                cout << "Capture from CAM " << inputName << " didn't work" << endl;
-            else
-                cout << "Capture from file " << vdofile << " failed" <<endl;
-            goto nocamera;
-        }
-
-        cout << "In capture ..." << endl;
-        for(int i = 0;; i++)
-        {
-            if( !capture.read(frame) )
-                break;
-
-            if (i == 0)
-            {
-                frame.copyTo( frame0 );
-                cvtColor(frame0, frame0Gray, COLOR_BGR2GRAY);
-            }
-            else
-            {
-                if (i%2 == 1)
-                {
-                    frame.copyTo(frame1);
-                    cvtColor(frame1, frame1Gray, COLOR_BGR2GRAY);
-                    ptr0 = frame0Gray;
-                    ptr1 = frame1Gray;
-                }
-                else
-                {
-                    frame.copyTo(frame0);
-                    cvtColor(frame0, frame0Gray, COLOR_BGR2GRAY);
-                    ptr0 = frame1Gray;
-                    ptr1 = frame0Gray;
-                }
-
-                alg->calc(ptr0, ptr1, flow);
-                split(flow, flow_vec);
-
-                if (i%2 == 1)
-                    frame1.copyTo(frameCopy);
-                else
-                    frame0.copyTo(frameCopy);
-                getFlowField(flow_vec[0].getMat(ACCESS_READ), flow_vec[1].getMat(ACCESS_READ), show_flow);
-                imshow("tvl1 optical flow field", show_flow);
-            }
-
-            char key = (char)waitKey(10);
-            if (key == 27)
-                break;
-            else if (key == 'm' || key == 'M')
-            {
-                ocl::setUseOpenCL(!cv::ocl::useOpenCL());
-                cout << "Switched to " << (ocl::useOpenCL() ? "OpenCL" : "CPU") << " mode\n";
-            }
-        }
-
-        capture.release();
-    }
-    else
-    {
-nocamera:
-        if (cmd.has("cpu_mode"))
-        {
-            ocl::setUseOpenCL(false);
-            std::cout << "OpenCL was disabled" << std::endl;
-        }
-        for(int i = 0; i <= LOOP_NUM; i ++)
-        {
-            cout << "loop" << i << endl;
-
-            if (i > 0) workBegin();
-
-            alg->calc(frame0, frame1, flow);
-            split(flow, flow_vec);
-
-            if (i > 0 && i <= LOOP_NUM)
-                workEnd();
-
-            if (i == LOOP_NUM)
-            {
-                if (useCPU)
-                    cout << "average CPU time (noCamera) : ";
-                else
-                    cout << "average GPU time (noCamera) : ";
-                cout << getTime() / LOOP_NUM << " ms" << endl;
-
-                getFlowField(flow_vec[0].getMat(ACCESS_READ), flow_vec[1].getMat(ACCESS_READ), show_flow);
-                imshow("PyrLK [Sparse]", show_flow);
-                imwrite(outpath, show_flow);
-            }
-        }
-    }
-
-    waitKey();
-
-    return EXIT_SUCCESS;
-}