Releases
1.5.5
saudet
released this
08 Mar 09:51
March 8, 2021 version 1.5.5
Bundle LLD executable in presets for LLVM as required by TVM on Windows
Prevent public static final
objects from getting deallocated by PointerScope
(issue bytedeco/javacv#1599 )
Fix compatibility of Leptonica with JavaFX by upgrading to libpng 1.6.37 (pull #1007 )
Introduce linux-arm64
build for CUDA, cuDNN, and NCCL (issue #735 )
Add new array constructors to opencv_core.Mat
that copy data for convenience (pull #1002 )
Rebase PrimitiveScalar
on PrimitiveScalarBase
in presets for Arrow for easy access to data()
(issue #998 )
Add NamedMetadataOperations.h
implementing data retrieval operations for LLVM nodes (pull #995 )
Enable OpenMP for ONNX Runtime on Mac once again (issue #917 )
Build OpenCV without OpenBLAS when environment variable NOOPENBLAS=yes
(pull #987 )
Enable OpenCL acceleration for DNNL (issue #938 )
Introduce monkey patching when loading presets for CPython to relocate home more reliably
Add display sample for librealsense2 (pull #978 )
Fix builds for libpostal on Mac and Windows (issue #903 )
Fix builds for NumPy and SciPy on Linux when using a cross compiler or not
Update presets for Visual Studio 2019 on Windows
Add presets for OpenCL 3.0, TVM 0.7.0 and bundle its official Java API (TVM4J) via the jnitvm_runtime
library
Include free()
in presets for FTTW as required by fftw_export_wisdom_to_string()
(issue bytedeco/javacpp#429 )
Include all missing header files from the opencv_ximgproc
module (issue #958 )
Disable assembly optimizations for libx264 with FFmpeg on Mac to work around crashes (issue bytedeco/javacv#1519 )
Add linux-armhf
and linux-arm64
builds for librealsense and librealsense2 (pull #951 )
License default builds of FFmpeg under LGPL v3 and move GPL-enabled builds to -gpl
extension (pull #950 )
Upgrade presets for OpenCV 4.5.1, FFmpeg 4.3.2, Arrow 3.0.0, Hyperscan 5.4.0, MKL 2021.1, OpenBLAS 0.3.13, ARPACK-NG 3.8.0, CMINPACK 1.3.8, FFTW 3.3.9, librealsense2 2.40.0 (pull #946 ), DNNL 2.1.1, CPython 3.9.2, NumPy 1.20.1, SciPy 1.6.1, Gym 0.18.0, LLVM 11.1.0 (pull #1001 ), OpenPose 1.7.0, CUDA 11.2.1, cuDNN 8.1.1, NCCL 2.8.4, MXNet 1.8.0, TensorFlow 1.15.5, TensorRT 7.2.3.4, ONNX 1.8.1, ONNX Runtime 1.7.0, Qt 5.15.2, Skia 2.80.2, cpu_features 0.6.0, and their dependencies
You can’t perform that action at this time.