Permalink
Commits on Dec 15, 2017
  1. OpenCV version++

    alalek committed Dec 15, 2017
    OpenCV 3.4.0-rc
  2. experimental version++

    alalek committed Dec 15, 2017
  3. cmake: fix ocv_target_compile_definitions() for world builds

    alalek committed Dec 15, 2017
    fixes CUDA build
  4. Merge pull request #10319 from catree:move_SimulatedAnnealingSolver_I…

    alalek committed Dec 15, 2017
    …mpl_cpp
    
    Move SimulatedAnnealingSolver::Impl in cpp file
  5. Merge pull request #10258 from savuor:fix/kmeans_channels

    savuor committed with alalek Dec 15, 2017
    * kmeans: number of channels in _centers fixed
    
    * fixedType() is checked now
  6. Merge pull request #10313 from alalek:rename_fix

    alalek committed Dec 15, 2017
    Fix file names
  7. Merge pull request #10287 from sturkmen72:update_createsamples_cpp

    sturkmen72 committed with alalek Dec 15, 2017
    * update createsamples
    
    adds command-line option -rngseed
    replaces rand() -> theRNG()
    
    * Update utility.cpp
    
    * apps(createsamples): fix warpPerspective pixels access bug
  8. Merge pull request #10315 from alalek:issue_10147_debug

    alalek committed Dec 15, 2017
    dnn: more debug information
  9. fix file names

    alalek committed Dec 15, 2017
  10. core:ppc Fix several issues for VSX (#10303)

    seiko2plus committed with vpisarev Dec 15, 2017
    - fix conversion intrinsics compatibility with xlc
    - implement odd-elements 2 to 4 conversion intrinsics
    - improve implementation of universal intrinsic v_popcount
    - rename FORCE_INLINE to VSX_FINLINE in vsx_utils.hpp
  11. Simulated Annealing for ANN_MLP training method (#10213)

    LaurentBerger committed with vpisarev Dec 15, 2017
    * Simulated Annealing for ANN_MLP training method
    
    * EXPECT_LT
    
    * just to test new data
    
    * manage RNG
    
    * Try again
    
    * Just run buildbot with new data
    
    * try to understand
    
    * Test layer
    
    * New data- new test
    
    * Force RNG in backprop
    
    * Use Impl to avoid virtual method
    
    * reset all weights
    
    * try to solve ABI
    
    * retry
    
    * ABI solved?
    
    * till problem with dynamic_cast
    
    * Something is wrong
    
    * Solved?
    
    * disable backprop test
    
    * remove ANN_MLP_ANNEALImpl
    
    * Disable weight in varmap
    
    * Add example for SimulatedAnnealing
  12. Merge pull request #10283 from zhijackchen:exr_export

    zhijackchen committed with alalek Dec 15, 2017
    * Fix issue #10114
    Convert table change
    From:
    CV_8U -> HALF
    CV_8S -> HALF
    CV_16U -> UINT
    CV_16S -> UINT
    CV_32S -> UINT
    CV_32F -> FLOAT
    To:
    CV_8U -> HALF
    CV_8S -> HALF
    CV_16U -> UINT
    CV_16S -> FLOAT
    CV_32S -> FLOAT loss precision
    CV_32F -> FLOAT
    Signed integer can't be presented well with UINT. Even adjust bias, CV16S and CV32S will be confused when load from exr file.
    Also fix CV_8S negative value incorrect bug
    
    * EXR import and export
    imread() from EXR returns CV_32F only
    imwrite() accepts CV_32 cv::Mat only and stores FLOAT images by default. Add imwrite() flag to store in HALF format.
    
    * fix compiling error
    
    * clean up
    
    * fix EXR import issues
  13. Faster-RCNN models support

    dkurt committed Dec 13, 2017
Commits on Dec 14, 2017
  1. core: remove raw SSE2/NEON implementation from convert.cpp (#9831)

    tomoaki0705 committed with vpisarev Dec 14, 2017
    * remove raw SSE2/NEON implementation from convert.cpp
      * remove raw implementation from Cvt_SIMD
      * remove raw implementation from cvtScale_SIMD
      * remove raw implementation from cvtScaleAbs_SIMD
      * remove duplicated implementation cvt_<float, short>
      * remove duplicated implementation cvtScale_<short, short, float>
      * add "from double" version of Cvt_SIMD
      * modify the condition of test ConvertScaleAbs
    
    * Update convert.cpp
    
    fixed crash in cvtScaleAbs(8s=>8u)
    
    * fixed compile error on Win32
    
    * fixed several test failures because of accuracy loss in cvtScale(int=>int)
    
    * fixed NEON implementation of v_cvt_f64(int=>double) intrinsic
    
    * another attempt to fix test failures
    
    * keep trying to fix the test failures and just introduced compile warnings
    
    * fixed one remaining test (subtractScalar)
  2. dnn: more debug information

    alalek committed Dec 14, 2017
  3. 3rdparty: update libtiff 4.0.9

    alalek committed Dec 8, 2017