* Add CUDA support for D3D11 interop. #13888
color_detail.hpp: fixed build error : dynamic initialization is not supported for a __constant__ variable.
directx.cpp: Add CUDA support(cl_nv_d3d11_sharing) for D3D11 interop. #13888
Update directx.cpp
Format adjustment.
Update directx.cpp
fix error.
Update directx.cpp
Format adjustment
Update directx.cpp
fix trailing whitespace.
fix format errors
convert indentation to spaces .
Trim trailing whitespace.
Add information about source of cl_d3d11_ext.h
Avoid unrelated changes.
Increase compile-time conditional judgment.
Increase the judgment of whether the OCL device has the required extensions at compile time.
Add compilation option `HAVE_CLNVEXT`.Check CL support in runtime.
Check result of `clGetExtensionFunctionAddressForPlatform` for KHR is invalid.It always can get the address(from OpenCL.dll),So I check NV support(from nvopencl64.dll) before KHR when `HAVE_CLNVEXT` is enabled.
Delete cl_d3d11_ext.h
Modified parameter list
fix "cannot open include file: 'CL/cl_d3d11_ext.h'"
remove not referenced var
fix C2143: syntax error
Improve compile-time judgment.
dlrectx.cpp Modify the detection order.
initializeContextFromD3D11Device:
```
// try with NV(Need to check it first)
// try with KHR
```
fix warnig C4100
Revert "fix warnig C4100"
This reverts commit 76e5becb67780071d0cbde61cc4f5f807ad7c5ac.
fix warning C4100
fix warning C4505
Format alignment
Format adjustment and automatically detect header files.
Automatically detect header files when users are not configured or configuration errors occur.
avoid unrelated changes.
Update .cmake
Update .cmake
* fix build errors
* fix warning:defined but not used
* Revert "fix warning:defined but not used"
This reverts commit 7ab3537cd0.
* fix warning:defined but not used
* fix build error for mac
* fix build error for win
* optimizing branch judgment
* Revert "optimizing branch judgment"
This reverts commit 88b72b870e.
* fix warning C4702: unreachable code
* remove unused code
* Fix problems that may lead to undefined behavior
* Add status check
* fix error C2664,C2665 : cannot convert argument
* Format adjustment
VSCODE will automatically format the indentation to 4 spaces in some situation.
* fix error C2440
* fix error C2440
* add cl_d3d11_ext.h
* Format adjustment
* remove unnecessary checks
When running with >1 OpenCV thread, KMeans index generation was
non-deterministic because of a RWW race. Issue is resolved by removing
the offending logic from the parallel section.
- allow cmake to check sanity of vsx aligned ld/st
- force universal intrinsics v_load_aligned/v_store_aligned
to failback to unaligned ld/st if cmake runtime vsx aligned test fail
* Fixed bug that made cuda::filter give corrupted output when different filters called in different threads. This was due to use of global texture variable shared by all filters.
* REV: fixed tab issue for opencv coding style...
The big endian data in the tiff_sample_data array was never being tested. This could be observed by e.g. changing the 9th byte in the big endian data from 0xde to something that should fail the test, e.g. 0xdd, and the test would still pass even though it should fail.
* Expose more C++ functionality in the Java wrapper of the Mat class
In particular expose methods for handling Mat with more than 2 dimensions
* add constructors taking an array of dimension sizes
* add constructor taking an existing Mat and an array of Ranges
* add override of the create method taking an array of dimension sizes
* add overrides of the ones and zeros methods taking an array of dimension sizes
* add override of the submat method taking an array of ranges
* add overrides of put and get taking arrays of indices
* add wrapper for copySize method
* fix crash in the JNI wrapper of the reshape(int cn, int[] newshape) method
* add test for each method added to Mat.java
* Fix broken test
* fix android lint error in BaseLoaderCallback (LongLogTag)
Lint Error that was causde by the to long TAG:
BaseLoaderCallback.java:31:
Error: The logging tag can be at most 23 characters,
was 31 (OpenCVLoader/BaseLoaderCallback) [LongLogTag]
Log.e(TAG, "Package installation failed!");
* add requested "OCV" prefix for android logging
Lab/XYZ modes have been postponed (color_lab.cpp):
- need to split code for tables initialization and for pixels processing first
- no significant performance improvements for switching between SSE42 / AVX2 code generation
According to the gstreamer docs [1], the GstMessage pointer returned by
gst_bus_pop() is nullable, meaning NULL is a valid return value.
Previously, gst_is_missing_plugin_message would throw an assert when its
message object parameter would fail the GST_IS_MESSAGE macro check,
crashing the entire process (unless running in a try-catch block of course).
Instead of relying on valid messages, check if the message object itself is
valid before passing it to other gstreamer functions.
[1] https://gstreamer.freedesktop.org/data/doc/gstreamer/head/gstreamer/html/GstBus.html#gst-bus-pop
Signed-off-by: Christopher N. Hesse <raymanfx@gmail.com>
* Enabled -Wnarrowing warning
* Fixed type narrowing issues
* Cast python constants
* Use long long for python constants
* Use int for python constants with fallback to long
* Update cv2.cpp
New computeECC function, and updated findTransformECC function to make gaussian filtering optional (#13837)
* fix for https://github.com/opencv/opencv/issues/12432 with doc and tests
* Added doc string for new parameter.
* Fixes suggested by Alalek for getting around ABI incompatibility.
* Update to docstring, to remove parameter that isn't relevant.
* More updates based on Alalek's usggestions.
Resize reworked using wide universal intrinsics (#13781)
* Added wide universal intrinsics optimized implementation for 3 channel bit-exact linear resize
* Reworked linear resize using new wide LUT intrinsics
* Fix for VSX intrinsics
All <arm_neon.h> includes in core/cv_cpu_dispatch.h are protected by an
ifndef __CUDACC__ to prevent attempting to use neon intrinsics when
compiling cuda kernels (.cu) -- this prevents hard errors such as
error: identifier "__builtin_neon_qi" is undefined
Add this same protection to flann/dist.h to fix compilation involving
flann.hpp.
objectPoints and imagePoints are not checked whether they're empty and
cause checkVector() to fail, thus result in a wrong error message.
Fixes: https://github.com/opencv/opencv/issues/6002
Due to size limit of shared memory, histogram is built on
the global memory for CV_16UC1 case.
The amount of memory needed for building histogram is:
65536 * 4byte = 256KB
and shared memory limit is 48KB typically.
Added test cases for CV_16UC1 and various clip limits.
Added perf tests for CV_16UC1 on both CPU and CUDA code.
There was also a bug in CV_8UC1 case when redistributing
"residual" clipped pixels. Adding the test case where clip
limit is 5.0 exposes this bug.
SVM sigmoid kernel fix (issue #13621) (#13718)
* Added test for sigmoid case for retrieving support vectors
* undo unhelpful test
* add test for sigmoid SVM with data that is easily separable into two concentric circles
* Update sigmoid kernel to use tanh(gamma * <x, y> + coef0) instead of -tanh(gamma * <x, y> + coef0)
* remove unnecessary constraint on coef0
* cleanup
* fixing inappropriate use of doubles
* Add f to float literal
* replace CV_Assert with ASSERT_EQ where appropriate
