The new dlib 19.3 Release has been announced for immediate availability. This release comes with a host of new features, improvements and bug fixes.
Dlib is a general purpose cross-platform open source C++ library. It is a cross-platform library written in ISO standard C++ and is officially supported on MS Windows, OS X, Linux, Solaris, the BSDs, and HP-UX.
This toolkit comes with machine learning algorithms and tools for creating complex software in C++. This allows the library to be used to solve real-world problems more efficiently. This has made the tool ideal for industry and academic use. Its use is widely seen in robotics, embedded devices, mobile phones and high-performance computing environments both and small and large.
Highlights of the dlib 19.3 Release
The biggest single new feature of this new release is the new capabilities in facial recognition. According to the release notes, it records 99.38% accuracy on the LFW benchmark.
There are performance improvements on the PowerPC machines. There is new support for loading of RGBA JPEG images as well as added support for Intel MKL FFT bindings.
There are a few bug fixes and improvements with MATLAB as well as bug fixes on the newer compilers. There are also usability improvements found in the DNN API and C++11 detection primarily on OS X.