Berlin, October 21st, 2019

Machine Learning with Dlib in Real Time

Kithara Software has announced the implementation of the software library Dlib into the Kithara real-time system. This way, the open library, which provides algorithms and tools for the programming of complex machine learning applications, can now be used in conjunction with hard real-time capabilities.

With Dlib, the Windows-based real-time operating system Kithara RealTime Suite for the first time supports dedicated software for the development in the field of machine learning. By employing real-time mechanisms, self-learning systems are able to utilize a more optimized communication of incoming information or trigger faster control reactions, such as for decision making.

“Dlib is not the first external function library to be integrated into our software. For years now, we made use of established image processing libraries with great success. With the direct connection of Dlib and Kithara RealTime Suite, we want to bridge real time and machine learning in order to provide easier access to such complex functions. This also allows for the practical usage of artificial intelligence with time-critical demands, such as for engineering”, Uwe Jesgarz, managing director of Kithara Software GmbH, explained.

Dlib is an open software library that can be implemented free of charge into any project. Besides algorithms for machine learning, including deep learning as well as other fields of application, the library also contains functions for machine vision. This means that real-time image capturing and processing, both of which are also provided by Kithara, can be combined with machine learning within the same real-time context.

For Kithara, the Dlib integration is the first direct move into real-time machine learning. An expansion in this field is already planned as further support for software solutions such as TensorFlow or MXNet are currently being evaluated.