Maksim Shabunin [Wed, 18 Jul 2018 12:24:58 +0000 (15:24 +0300)]
Added more strict checks for empty inputs to compare, meanStdDev and RNG::fill
Vadim Pisarevsky [Thu, 26 Jul 2018 09:09:35 +0000 (09:09 +0000)]
Merge pull request #12052 from dkurt:dnn_ie_torch_tests
Vadim Pisarevsky [Thu, 26 Jul 2018 09:04:28 +0000 (12:04 +0300)]
further improvements in split & merge; started using non-temporary store instructions (#12063)
* 1. changed static const __m128/256 to const __m128/256 to avoid wierd instructions and calls inserted by compiler.
2. added universal intrinsics that wrap MOVNTPS and other such (non-temporary or "no cache" store) instructions. v_store_interleave() and v_store() got respective flags/overloaded variants
3. rewrote split & merge to use the "no cache" store instructions. It resulted in dramatic performance improvement when processing big arrays
* hopefully, fixed some test failures where 4-channel v_store_interleave() is used
* added missing implementation of the new universal intrinsics (v_store_aligned_nocache() etc.)
* fixed silly typo in the new intrinsics in intrin_vsx.hpp
* still trying to fix VSX compiler errors
* still trying to fix VSX compiler errors
* still trying to fix VSX compiler errors
* still trying to fix VSX compiler errors
Dmitry Kurtaev [Tue, 24 Jul 2018 16:12:58 +0000 (19:12 +0300)]
Faster-RCNN anf RFCN models on CPU using Intel's Inference Engine backend.
Enable Torch layers tests with Intel's Inference Engine backend.
Alexander Alekhin [Tue, 24 Jul 2018 19:45:27 +0000 (19:45 +0000)]
Merge pull request #12048 from mshabunin:fix-static-2
Alexander Alekhin [Tue, 24 Jul 2018 19:44:25 +0000 (19:44 +0000)]
Merge pull request #12047 from alalek:issue_12045
Alexander Alekhin [Tue, 24 Jul 2018 19:43:49 +0000 (19:43 +0000)]
Merge pull request #12020 from alalek:videoio_backends_query_api
Vadim Pisarevsky [Tue, 24 Jul 2018 14:27:56 +0000 (17:27 +0300)]
converted split() & merge() to wide univ intrinsics (#12044)
* fixed/updated v_load_deinterleave and v_store_interleave intrinsics; modified split() and merge() functions to use those intrinsics
* fixed a few compile errors and bug in v_load_deinterleave(ptr, v_uint32x4& a, v_uint32x4& b)
* fixed few more compile errors
Maksim Shabunin [Tue, 24 Jul 2018 11:14:13 +0000 (14:14 +0300)]
More issues found by static analysis
Alexander Alekhin [Tue, 24 Jul 2018 13:03:41 +0000 (13:03 +0000)]
Merge pull request #12021 from dkurt:dnn_ie_tf_ssd
Alexander Alekhin [Tue, 24 Jul 2018 12:34:04 +0000 (12:34 +0000)]
Merge pull request #12037 from dkurt:test_openvino_models
Alexander Alekhin [Tue, 24 Jul 2018 11:10:01 +0000 (11:10 +0000)]
Merge pull request #12041 from mshabunin:fix-static-2
Alexander Alekhin [Tue, 24 Jul 2018 10:08:58 +0000 (13:08 +0300)]
objdetect(qr): update test code
improve error checks
Alexander Alekhin [Tue, 24 Jul 2018 08:17:20 +0000 (08:17 +0000)]
Merge pull request #12042 from dkurt:fix_opencv_as_submodule
Alexander Alekhin [Mon, 23 Jul 2018 15:17:51 +0000 (18:17 +0300)]
dnn: update tests for OpenVINO models
Alexander Alekhin [Mon, 23 Jul 2018 14:58:10 +0000 (17:58 +0300)]
ts: add findDataDirectory() function
Dmitry Kurtaev [Mon, 23 Jul 2018 20:34:50 +0000 (23:34 +0300)]
Do not copy cv_cpu_helper.h to parent if OpenCV is a submodule
Dmitry Kurtaev [Fri, 20 Jul 2018 15:58:37 +0000 (18:58 +0300)]
Add a sample which tests OpenVINO models
Maksim Shabunin [Mon, 23 Jul 2018 14:22:47 +0000 (17:22 +0300)]
Fixed several issues found by static analysis tools
Alexander Alekhin [Mon, 23 Jul 2018 07:52:42 +0000 (07:52 +0000)]
Merge pull request #12031 from alalek:dnn_ocl_eliminate_getUMat_with_bad_lifetime
Paul92 [Sun, 22 Jul 2018 14:08:29 +0000 (15:08 +0100)]
Merge pull request #12032 from Paul92:mser-sample-improvments
Mser sample improvments (#12032)
* Fixed bug in detect_mser sample
Wrong number of colors used to generate the synthetic images
* Formatting improvements
* Using safer casts
* Improved readability of legend generation
* Various readability fixes in detect_mser sample
Alexander Alekhin [Sat, 21 Jul 2018 03:50:58 +0000 (03:50 +0000)]
Merge pull request #12016 from tyl12:fix_videoio_static_para
Alexander Alekhin [Sat, 21 Jul 2018 03:30:55 +0000 (03:30 +0000)]
Merge pull request #12003 from mshabunin:avi-error-messages
Teng Yiliang [Thu, 19 Jul 2018 08:40:27 +0000 (16:40 +0800)]
use struct member width_set/height_set to replace static width/height.
the static variables will cause race-condition when operating in
multithread scenarios.
Signed-off-by: Teng Yiliang <ylteng@outlook.com>
Signed-off-by: Teng Yiliang <yiliang.teng@weimob.com>
Alexander Alekhin [Fri, 20 Jul 2018 14:26:35 +0000 (17:26 +0300)]
dnn(ocl): don't use getUMat() for long live objects
Alexander Alekhin [Fri, 20 Jul 2018 11:27:27 +0000 (11:27 +0000)]
Merge pull request #12029 from tomoaki0705:fixBuildVS2013BinarySuffix
Tomoaki Teshima [Fri, 20 Jul 2018 09:09:17 +0000 (18:09 +0900)]
fix build error on Visual Studio 2013
* replace binary literal prefix to hexadecimal literal prefix
Alexander Alekhin [Thu, 19 Jul 2018 20:50:42 +0000 (20:50 +0000)]
Merge pull request #12022 from mshabunin:fix-ie-build
Maksim Shabunin [Thu, 19 Jul 2018 16:22:23 +0000 (19:22 +0300)]
dnn, IE backend: updated to match new interface
Alexander Alekhin [Thu, 19 Jul 2018 15:50:33 +0000 (15:50 +0000)]
Merge pull request #12019 from mshabunin:static-debug-assert
Alexander Alekhin [Thu, 19 Jul 2018 14:38:46 +0000 (14:38 +0000)]
Merge pull request #11781 from dkurt:dnn_uint8_inputs
Alexander Alekhin [Thu, 19 Jul 2018 13:14:50 +0000 (16:14 +0300)]
videoio: add routines to query information about backends API
into cv::videoio_registry namespace
Dmitry Kurtaev [Thu, 19 Jul 2018 14:05:56 +0000 (17:05 +0300)]
Run entire SSDs from TensorFlow using Intel's Inference Engine
Alexander Alekhin [Thu, 19 Jul 2018 12:09:48 +0000 (12:09 +0000)]
Merge pull request #12007 from alalek:fix_ts_perf_test_macro
Dmitry Kurtaev [Mon, 4 Jun 2018 20:51:28 +0000 (23:51 +0300)]
uint8 inputs for deep learning networks
Alexander Alekhin [Wed, 18 Jul 2018 12:20:43 +0000 (15:20 +0300)]
ts: fix PERF_TEST() macro to allow test_case name reusing
Example (reuse 'Transform' test case):
PERF_TEST(Transform, getPerspectiveTransform_1000) { ... }
PERF_TEST(Transform, getPerspectiveTransform_QR_1000) { ... }
Alexander Alekhin [Thu, 19 Jul 2018 09:37:36 +0000 (09:37 +0000)]
Merge pull request #12008 from alalek:test_remove_verbose_messages
Alexander Alekhin [Wed, 18 Jul 2018 13:09:27 +0000 (16:09 +0300)]
core(test): remove verbose messages
Maksim Shabunin [Wed, 18 Jul 2018 12:53:16 +0000 (15:53 +0300)]
Enable debug assertions for static analysis builds
Maksim Shabunin [Wed, 18 Jul 2018 12:22:42 +0000 (15:22 +0300)]
AVI container: verbose error messages
Alexander Alekhin [Tue, 17 Jul 2018 16:24:05 +0000 (16:24 +0000)]
Merge pull request #11104 from asciian:reading_from_stream
Alexander Alekhin [Tue, 17 Jul 2018 15:41:36 +0000 (15:41 +0000)]
Merge pull request #11986 from alalek:build_eliminate_gcc8_warnings
Alexander Alekhin [Tue, 17 Jul 2018 15:40:48 +0000 (15:40 +0000)]
Merge pull request #11984 from mshabunin:fix-static-1
Alexander Alekhin [Tue, 17 Jul 2018 15:38:46 +0000 (15:38 +0000)]
Merge pull request #11968 from allnes:detect_qr_code
Alexander Alekhin [Tue, 17 Jul 2018 15:35:07 +0000 (15:35 +0000)]
Merge pull request #11964 from mshabunin:avi-container-fixes
Alexander Alekhin [Tue, 17 Jul 2018 15:34:08 +0000 (15:34 +0000)]
Merge pull request #11990 from alalek:clone_nodiscard_attribute
Alexander Alekhin [Tue, 17 Jul 2018 15:14:41 +0000 (15:14 +0000)]
Merge pull request #11994 from alalek:cmake_find_package_inference_engine
Maksim Shabunin [Tue, 17 Jul 2018 14:50:50 +0000 (17:50 +0300)]
Check for empty Mat in compare, operator= and RNG::fill, fixed related tests
Maksim Shabunin [Tue, 17 Jul 2018 13:14:54 +0000 (16:14 +0300)]
Fixed issues found by static analysis (mostly DBZ)
Alexander Alekhin [Tue, 17 Jul 2018 12:30:56 +0000 (15:30 +0300)]
cmake: prefer using find_package(InferenceEngine)
Alexander Alekhin [Tue, 17 Jul 2018 11:20:02 +0000 (11:20 +0000)]
Merge pull request #11959 from pengli:3.4
Dmitry Kurtaev [Tue, 17 Jul 2018 10:41:46 +0000 (13:41 +0300)]
Add Java overloads for each default argument (#11940)
* Add Java overloads for each default argument
* Add "fisheye_" prefix for cv::fisheye:: functions and enums
Alexander Alekhin [Tue, 17 Jul 2018 10:27:21 +0000 (10:27 +0000)]
Merge pull request #11989 from csukuangfj:patch_2
Li Peng [Thu, 12 Jul 2018 07:16:32 +0000 (15:16 +0800)]
enable concat layer fuse for OCL target
Signed-off-by: Li Peng <peng.li@intel.com>
Vadim Pisarevsky [Mon, 16 Jul 2018 15:57:24 +0000 (18:57 +0300)]
Wide univ intrinsics (#11953)
* core:OE-27 prepare universal intrinsics to expand (#11022)
* core:OE-27 prepare universal intrinsics to expand (#11022)
* core: Add universal intrinsics for AVX2
* updated implementation of wide univ. intrinsics; converted several OpenCV HAL functions: sqrt, invsqrt, magnitude, phase, exp to the wide universal intrinsics.
* converted log to universal intrinsics; cleaned up the code a bit; added v_lut_deinterleave intrinsics.
* core: Add universal intrinsics for AVX2
* fixed multiple compile errors
* fixed many more compile errors and hopefully some test failures
* fixed some more compile errors
* temporarily disabled IPP to debug exp & log; hopefully fixed Doxygen complains
* fixed some more compile errors
* fixed v_store(short*, v_float16&) signatures
* trying to fix the test failures on Linux
* fixed some issues found by alalek
* restored IPP optimization after the patch with AVX wide intrinsics has been properly tested
* restored IPP optimization after the patch with AVX wide intrinsics has been properly tested
Alexander Alekhin [Mon, 16 Jul 2018 15:54:34 +0000 (15:54 +0000)]
Merge pull request #11957 from alalek:issue_11956
Alexander Alekhin [Mon, 16 Jul 2018 15:51:26 +0000 (15:51 +0000)]
Merge pull request #11970 from dkurt:dnn_enable_tf_tests
Alexander Alekhin [Mon, 16 Jul 2018 15:49:19 +0000 (15:49 +0000)]
Merge pull request #11985 from alalek:cmake_update_inference_engine_detection
Alexander Alekhin [Mon, 16 Jul 2018 14:59:24 +0000 (17:59 +0300)]
core: CV_NODISCARD macro with semantic of [[nodiscard]] attr
[[nodiscard]] is defined in C++17.
There is fallback alias for modern GCC / Clang compilers.
Kuang Fangjun [Sat, 14 Jul 2018 08:14:35 +0000 (16:14 +0800)]
remove a useless statement.
Alexander Alekhin [Mon, 16 Jul 2018 12:34:59 +0000 (15:34 +0300)]
build: eliminate GCC8 warnings
Alexander Alekhin [Mon, 16 Jul 2018 14:18:12 +0000 (14:18 +0000)]
Merge pull request #11969 from alalek:core_Matx_inv_solve_templates
Alexander Alekhin [Mon, 16 Jul 2018 13:41:45 +0000 (13:41 +0000)]
Merge pull request #11951 from alalek:cuda_use_external_cmake_module
Alexander Alekhin [Mon, 16 Jul 2018 11:10:32 +0000 (11:10 +0000)]
Merge pull request #11963 from dkurt:dnn_cl_fix_matmul
Alexander Alekhin [Mon, 16 Jul 2018 10:58:59 +0000 (13:58 +0300)]
cmake: update DL IE detection to align with OpenVINO 2018 R2
These CMake options are enough:
- `-DWITH_INF_ENGINE=ON`
- `-DIE_PLUGINS_PATH=lib/ubuntu_16.04/intel64`
- `-DENABLE_CXX11=ON` (OpenCV 3.4 only)
Maksim Shabunin [Fri, 13 Jul 2018 11:07:30 +0000 (14:07 +0300)]
AVI container: use C++ streams for file operations, check some operations for overflow
Alexander Alekhin [Mon, 16 Jul 2018 09:48:33 +0000 (09:48 +0000)]
Merge pull request #11965 from alalek:issue_11944
Alexander Alekhin [Mon, 16 Jul 2018 09:24:32 +0000 (09:24 +0000)]
Merge pull request #11967 from catree:add_tutorial_ml_java_python
Alexander Alekhin [Sun, 15 Jul 2018 15:06:16 +0000 (15:06 +0000)]
Merge pull request #11971 from csukuangfj:patch_1
Alexander Nesterov [Fri, 13 Jul 2018 21:51:04 +0000 (18:51 -0300)]
Update qrcode algorithm: fix seg. fault with kmeans
Kuang Fangjun [Sat, 14 Jul 2018 08:05:29 +0000 (16:05 +0800)]
avoid negative index.
Alexander Alekhin [Fri, 13 Jul 2018 15:52:20 +0000 (18:52 +0300)]
core: eliminate 'if' logic from Matx::inv()/solve()
- 'if' logic is moved into templates.
- removed unnecessary cv::Mat objects creation.
- fixed inv() test (invA * A == eye)
- added more Matx tests to cover all defined template specializations
Dmitry Kurtaev [Fri, 13 Jul 2018 16:47:50 +0000 (19:47 +0300)]
Enable TensorFlow networks tests for different backends and targets
Alexander Alekhin [Fri, 13 Jul 2018 15:15:03 +0000 (15:15 +0000)]
Merge pull request #11955 from terfendail:matx_solve_fix2
Vadim Pisarevsky [Fri, 13 Jul 2018 14:26:17 +0000 (14:26 +0000)]
Merge pull request #11921 from allnes:detect_qr_code
Alexander Alekhin [Fri, 13 Jul 2018 14:25:42 +0000 (14:25 +0000)]
Merge pull request #11966 from alalek:core_solve_method_check
catree [Fri, 13 Jul 2018 13:11:00 +0000 (15:11 +0200)]
Add Java and Python code for ML tutorials.
Dmitry Kurtaev [Fri, 13 Jul 2018 10:23:45 +0000 (13:23 +0300)]
Fix fully-connected layer in case of number of rows less than 4
catree [Fri, 13 Jul 2018 13:05:54 +0000 (15:05 +0200)]
Add overloaded cv::PCACompute() that returns also the eigenvalues. Useful for Java and Python OpenCV where PCA is not available.
Alexander Alekhin [Fri, 13 Jul 2018 11:57:46 +0000 (14:57 +0300)]
imgproc(getPerspectiveTransform): add configuration parameter
Alexander Alekhin [Fri, 13 Jul 2018 11:39:32 +0000 (14:39 +0300)]
imgproc(getPerspectiveTransform): solve(DECOMP_SVD -> LU)
Alexander Alekhin [Fri, 13 Jul 2018 11:34:46 +0000 (14:34 +0300)]
imgproc(perf): update getPerspectiveTransform perf test
Function is very fast, so 0.000 ms results are useless.
1000 runs requires 25ms on i7-6700K.
Alexander Alekhin [Fri, 13 Jul 2018 12:29:47 +0000 (12:29 +0000)]
Merge pull request #11960 from dkurt:dnn_cl_clip_kernel
Alexander Alekhin [Fri, 13 Jul 2018 12:11:57 +0000 (15:11 +0300)]
core: solve(): add check for passed 'method' values
Dmitry Kurtaev [Fri, 13 Jul 2018 07:48:31 +0000 (10:48 +0300)]
Clip kernel for OpenCL PriorBox layer
Alexander Nesterov [Tue, 10 Jul 2018 17:24:09 +0000 (14:24 -0300)]
Update detect QRCode algorithm
Alexander Alekhin [Thu, 12 Jul 2018 17:14:23 +0000 (20:14 +0300)]
core: use "explicit" for Matx() ctor
Vitaly Tuzov [Thu, 12 Jul 2018 16:19:51 +0000 (19:19 +0300)]
Fixed unreachable code warnings for Matx::solve()
Alexander Alekhin [Wed, 11 Jul 2018 10:47:56 +0000 (13:47 +0300)]
CUDA: allow to use external FindCUDA from modern CMake
CMake 3.9.0+ is required
Alexander Alekhin [Thu, 12 Jul 2018 09:48:08 +0000 (09:48 +0000)]
Merge pull request #11949 from berak:dnn_object_detection_typo
berak [Thu, 12 Jul 2018 09:06:53 +0000 (11:06 +0200)]
dnn: fix typo in object_detection.cpp sample
Alexander Alekhin [Thu, 12 Jul 2018 07:37:55 +0000 (07:37 +0000)]
Merge pull request #11941 from alalek:dnn_ocl_fix_verify_umat_mapping
Alexander Alekhin [Wed, 11 Jul 2018 20:27:17 +0000 (20:27 +0000)]
Merge pull request #11942 from catree:add_tutorial_core_java_python
Vitaly Tuzov [Wed, 11 Jul 2018 19:00:57 +0000 (22:00 +0300)]
Merge pull request #11904 from terfendail/matx_solve_fix
Fixed Matx::solve function for non-square matrixes (#11904)
catree [Wed, 11 Jul 2018 17:38:16 +0000 (19:38 +0200)]
Add Java and Python code for the following tutorials:
- Changing the contrast and brightness of an image!
- Operations with images
Alexander Alekhin [Wed, 11 Jul 2018 16:11:36 +0000 (19:11 +0300)]
dnn(ocl): fix wrong usage of stalled .getMat() pointers
Temporary object lifetime must be greater than pointer usage.
Dmitry Kurtaev [Wed, 11 Jul 2018 09:48:34 +0000 (12:48 +0300)]
Replace std::vector<char> to std::vector<uchar> for Java bindings of dnn importers
Alexander Alekhin [Wed, 11 Jul 2018 12:01:31 +0000 (12:01 +0000)]
Merge pull request #11936 from berak:dnn_shufflelayer_name
Alexander Alekhin [Wed, 11 Jul 2018 09:33:24 +0000 (09:33 +0000)]
Merge pull request #11927 from pengli:3.4
Li Peng [Tue, 10 Jul 2018 04:43:03 +0000 (12:43 +0800)]
Fix gemmlike convolution input reading
use vload3 for half3 or float3 input vector reading,
also check read position to see if it exceed input width
Signed-off-by: Li Peng <peng.li@intel.com>