The main objective of work package 5 is to translate and validate the different nonlocal algorithms developed in work package 4 on real world problems. In particular, we want to use nonlocal methods for high-resolution 3D/4D biomedical data and massive 3D point cloud data in order to evaluate their potential. On this wide range of data sets we will tackle challenging tasks including denoising, segmentation, and classification problems. Furthermore, we will study general matrix recovery problems, which naturally arise in a variety of practical applications, e.g., recommendation systems. The success of these methods has already been ascertained in several cases. Our aim is to continue and expand upon these early results with particular focus on the aforementioned applications. We need to acquire and provide the needed test data sets for the evaluation of nonlocal methods within this project by close collaborations with the participating industrial partners. We aim to develop tailored software solutions together with the partner companies, which exploit the potential of nonlocal methods and increase the competitiveness of the companies.