OpenCV 3.2 Computer Vision Stack Released

A big update to the Open-Source Computer Vision library was quietly released just before Christmas.

OpenCV 3.2 features many improvements to the DNN module, support for more image formats and camera back-ends, an interactive camera calibration application, more algorithms have been added, support for the newest operating systems, various Intel and ARM architecture optimizations, support for using a vendor-provided OpenVX and LAPACK/BLAS libraries, and much more.

Thanks to the Google Summer of Code are also some improvements and documentation enhancements, including OpenCV tutorials in Python / C++ / Java. OpenCV 3.2 also now exposes OpenCL acceleration to Python, the DNN module has refactored OpenCL acceleration, several OpenCL kernels were added that are optimized for Intel graphics, CUDA 8 is now supported, and a lot of other changes.

More details on the OpenCV 3.2 release via the announcement at OpenCV.org while a closer look at the changes for this Computer Vision library update can be found from this GitHub page.