| ID | Branch | Title and description | Win | Lin | Lin2 | Mac | And | Doc | Author | Assigned to | Merge status |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 900 | master | Restore "Split GPU module into several submodules" |
3574 success |
3596 success |
1112 success |
3482 success |
2492 success |
2562 success |
jet47 | Nobody | Under review |
| 899 | master | Fix drawing documentation referring to C API macros |
3573 success |
3595 success |
1111 success |
3481 success |
2491 success |
2561 warnings |
LeszekSwirski | vpisarev | Unsuccessful builds |
| 895 | 2.4 | ocl perf verifies results while running performance tests add more test cases. Fix a bug may lead hunging. |
3577 success |
3599 success |
1115 success |
3485 success |
2495 success |
2565 success |
bitwangyaoyao | vpisarev | Under review |
| 893 | 2.4 | Fixed some warnings and errors on arm platform (GPU module) |
3565 success |
3587 warnings |
1103 success |
3473 success |
2483 success |
2553 success |
jet47 | Nerei | Unsuccessful builds |
| 891 | 2.4 | Bug fix: corrected a bug in HOGDescriptor::detectMultiScale() The detectMultiScale did not group weights along with ROIs. A new protected method is added to HOGDescriptor structure to handle weight grouping, instead of using the generic groupRectangles() function in cascadedetect.cpp. A class definition had to be moved out of cascadedetect,cpp into cascadedetect.hpp to accommodate the change. Fixes issue #3021 (http://code.opencv.org/issues/3021) |
3579 success |
3601 success |
1117 success |
3487 success |
2497 success |
2567 success |
NCBee | cuda-geek | Under review |
| 890 | master | fix a bug(DetectorType never change) changed line218 -> line220 Presentation: because line 220 give the global variable mDetectorType and in line 230 it will be compared with mDetectorType !!! it will never be unequal ~ fix: change mDetectorType(previous global var) to a new local var tmpDetectorType |
3567 success |
3589 success |
1105 success |
3475 success |
2485 success |
2555 success |
caorong | kirill-kornyakov | Under review |
| 889 | master | Fixed include name in OpenCL on OS X The filename is actually ``opencl.h`` in ``/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.8.sdk/System/Library/Frameworks/OpenCL.framework/Versions/A/Headers``, which causes build failures with case-sensitive filesystem. |
3568 success |
3590 success |
1106 success |
3476 success |
2486 success |
2556 success |
janm399 | vpisarev | Under review |
| 888 | master | Fix CMake MSVC64 setting Workaround for CMake bug (http://www.cmake.org/Bug/view.php?id=13896) |
3562 success |
3584 success |
1100 success |
3470 success |
2480 success |
2550 success |
LeszekSwirski | vpisarev | Under review |
| 887 | master | Fix visibility of g_linked_ptr_mutex on Windows Reverts some changes in b8ed00bd642971877e1dfda6e0815d7211a7ace2 |
3569 success |
3591 success |
1107 success |
3477 success |
2487 success |
2557 success |
LeszekSwirski | olegsklyarov | Under review |
| 882 | 2.4 | Copy ocl::queryDeviceInfo interface from master to 2.4. Affected functions surf.ocl, pyrlk.ocl and hog.ocl are updated with the change. |
3556 success |
3578 success |
1094 success |
3464 success |
2474 success |
2544 success |
pengx17 | vpisarev | Under review |
| 876 | 2.4 | 2.4 Adding BackgroundSubtractorMOG2 to python binding |
3570 success |
3592 success |
1108 success |
3478 success |
2488 success |
2558 success |
noskill | vpisarev | Under review |
| 870 | master | Build Opencv with Qt5 using -DWITH_QT5=ON ...tested (I need to figure out where the sample files are). Includes a new HAVE_QT5 define that is very similar to the original HAVE_QT define except in window_QT.h where it is used to include the qt5 widget headers instead of the qt4 gui headers. Note: https://qt-project.org/forums/viewthread/23459/ explains why window_QT.cpp line 3103 checks for QT_OPENGL_ES_2. GL_PERSPECTIVE_CORRECTION_HINT is not defined in gl2.h. If someone knows the equivalent call in opengl es2 it should probably be included. |
3526 success |
3548 success |
1064 success |
3434 success |
2444 success |
2514 success |
bpmckinnon | vpisarev | Under review |
| 868 | master | Fixed crash Without that line, the sample crashes on my system with the following error: OpenCV Error: OpenGL API call (Can't load OpenGL extension [glGenerateMipmap]) in unknown function, file ......opencv-2.4.5modulescoresrcgl_core_3_1.cpp, line 141 Traceback (most recent call last): File "C:facedetect.py", line 60, in |
3480 success |
3502 success |
1018 success |
3388 success |
2398 success |
2468 success |
Pastafarianist | vpisarev | Under review |
| 863 | master | Ubuntu compile Just some changes to allow compile on Ubuntu for limits and threads. |
3575 success |
3597 running |
1113 success |
3483 success |
2493 success |
2563 success |
gunnjo | vpisarev | Waiting for builds |
| 856 | master | Fix for spurious assert when comparing multi-channel histograms When comparing histograms that look like multi-channel images (e.g a 3D histogram, of 4x4x4 bins, might appear as a CV_32FC4 matrix), cv::compareHist would complain because it was expecting the matrix type() == CV_32F. Now we test for matrix depth() == CV_32F instead. This appeared to be a particular problem when using the Python API - presumably a 3D `i x j x k` element `numpy.ndarray` was getting automagically converted to a 2D `i x j` image with `k` channels. |
3576 success |
3598 success |
1114 warnings |
3484 success |
2494 success |
2564 success |
ericgreveson | vpisarev | Unsuccessful builds |
| 853 | 2.4 | Bugs: After scaling back to original image, some detected ROI will outside the original image ROI May be a bug in modules/objdetect/src/cascadedetect.cpp. |
3578 success |
3600 success |
1116 success |
3486 success |
2496 success |
2566 success |
icylord | cuda-geek | Under review |
| 837 | master | Add feature #2615 : ocl::imshow display oclMat using cl_khr_gl_sharing extension Add new function ocl::imshow( const String& winname, const oclMat& _img ) to ocl.hpp A hidden window is created to generate OpenGL context, from which the OpenCL context is created. Only Windows system is supported. |
3571 fail |
3593 fail |
1109 fail |
3479 fail |
2489 fail |
2559 fail |
hewj03 | apavlenko | Unsuccessful builds |
| 830 | master | Add implementations for cvSetMemoryManager and some related APIs. Allow users to activate memory tool at runtime. Some descriptions: 1. Provide the implementation for cvSetMemoryManager(). cvRemoveMemoryManager() is also added but it just simply switches the memory manager to the default allocator/deallocator, i.e. system malloc/free. Currently the code limits the number of user-defined memory managers to 1022. 2. Allow users to activate memory pool at runtime. Currently in opencv, if we want to activate the memory pool, we need to remove "#define CV_USE_SYSTEM_MALLOC" in alloc.c and then recompile opencv. In this patch, we can activate the memory tool by involving cvTurnOnMemoryPool(). Also we can turn off it via cvTurnOffMemoryPool(). When calling cvTurnOnMemoryPool(), users can specific the block size used in the memory pool (if an object's size is larger than block size, it will be allocated by system malloc/free). However currently it doesn't allow to change the block size. That is to say, if the memory pool is turned on, it can only be re-turned on with the same block size. 3. The code doesn't allow user to apply different memory manager to different threads. If this feature is needed, we may need to change lots of code in opencv's basic structure and may result in lots of problems. |
3429 success |
3451 success |
967 success |
3337 success |
2347 success |
2417 success |
ArtanisCV | vpisarev | Under review |
| 797 | master | fix installation of main module headers see: http://code.opencv.org/issues/2897 The order of the 2 conditions is really important, as every ```MATCHES``` operator sets ```${CMAKE_MATCH_1}```, which is needed inside the condition block. Ideally, this should be done using a single regex with a negative forward lookahead, but CMake doesn't seem to support this... |
3489 success |
3511 success |
1027 warnings |
3397 success |
2407 success |
2477 success |
vhdirk | vpisarev | Unsuccessful builds |
| 773 | 2.4 | changed contacts in tutorual |
3555 success |
3577 success |
1093 success |
3463 success |
2473 success |
2543 success |
alekcac | alekcac | Under review |
| 733 | master | Add ANN C langrage example Add CvANN_MLP C example. The ANN is used to calculate the average of two number. |
3330 success |
3350 success |
866 success |
3236 success |
2246 success |
2316 success |
godspeed1989 | kirill-kornyakov | Under review |
| 696 | master | added compute function to Feature2D (dirty workaround for #2699) This is an ugly attempt to fix #2699. Currently it seems to be impossible to run any feature descriptor on user-provided keypoints from python. Not sure if this solution should be merged in it's current form. |
3276 success |
3296 success |
812 success |
3182 success |
2192 success |
2261 success |
znah | vpisarev | Under review |
| 674 | master | Bugfix to potential inaccuracy in function:polyfit 1.Modify the default datatype(CV_32F) to CV_64F. 2.Re-write the generation of X matrix,reducing the times of multiplication. 3.Add a bool argument to control output datatype(float or double) The problem has been reported at http://code.opencv.org/issues/2887 |
3474 success |
3496 success |
1012 success |
3382 success |
2392 success |
2462 success |
chouclee | vpisarev | Under review |
| 604 | master | Mono / .NET bindings hi, using IKVM (http://www.ikvm.net/) i've created bindings for Mono / .NET using the existing java bindings. in order to get them built, IKVM needs to be in the system's PATH environment variable. i modified the java for desktop bindings a little so they do the System.loadLibrary call on their own. doing it the same way as in the java examples (letting the user call it) doesn't work in .NET due to different class loading behaviour. i'm fairly new to CMake, so any hints/help in that direction are appreciated - for example i've no clue what'd be the proper way to copy native dependencies (ikvm-native-win32) based on the current target platform (x86/x64). i'd like to see an output folder like this: mono/opencv-mono.dll -- the mono bindings mono/IKVM.*.dll -- managed assemblies needed by opencv-mono.dll mono/x86/opencv_java249.dll -- native dependency needed by opencv-mono.dll mono/x86/ikvm-native-win32-x86.dll -- native dependency needed by IKVM.Runtime.JNI.dll mono/x64/... i've not yet written any tests yet. simply tried some basic examples which worked just fine. wanted to sort out above mentioned build issues first. thanks, elias |
3310 fail |
3330 fail |
846 fail |
3216 fail |
2226 fail |
2296 fail |
azeno | apavlenko | Unsuccessful builds |
| 406 | master | Fixed DynamicAdaptedFeatureDetector (#1334) |
3042 fail |
3055 fail |
563 fail |
2932 fail |
1942 fail |
2014 fail |
ilysenkov | vpisarev | Unsuccessful builds |