A machine learning based approach for value stream map digitization

The Value Stream Mapping method provides a comprehensive overview of a process, pinpoints inefficiencies, and thereby offers a roadmap for eliminating waste and enhancing added value. Simplifications are often made enabling manual evaluations but at the same time sacrificing valuable information about dynamic process behavior. Simulation software enables the aspect of dynamics to be considered. This software requires a digital representation of the value stream map created during Value Stream Mapping, which is typically created in non-digital formats unusable in software. This paper presents an approach that enables the automated digitization of such non-digital value stream maps.