KIRAS – K.REX

01.12.2019 – 30.11.2021

Investigators in (criminal) prosecution have to evaluate huge and ever increasing amounts of heterogeneous data and documents. The need for resources for computer-based, analytical support is therefore urgent. However, the calculation tools for text analysis are not sufficient.  For example, the intuitive ability to draw conclusions from the structure of documents or to extract optical features, pictorial elements (such as logos, stamps or written additions) and their positions with high reliability are missing. K.Rex proposes an approach to capture elements of human perception in machine-learned models.