Docker support for instance segmentation

arivis Pro 4.2 introduced the possibility of performing instance segmentation using Deep Learning. This DL segmentation tool relies on various layers of dependent libraries. Rather than forcing configurations that could cause a conflict with other softwares, arivis Pro uses Docker technology to embed all the necessary dependencies for these DL models into a Docker container. A Docker container has all the information required to execute the segmentation sandboxed in such a way as to be independent of the system configuration and without risking conflicts with other software installations.

Virtualization features in both the BIOS (on bear-metal installations) and Operating System are required for containers to work. Typically Intel Virtualization Technology (VT-x) or SVM Mode (AMD-V) on AMD systems, see https://docs.docker.com/desktop/troubleshoot-and-support/troubleshoot/topics/#virtualization

Only NVIDIA drivers are currently supported for CUDA processing.

For Docker Desktop, WSL2 must be used for Paravirtualization support, see https://docs.docker.com/desktop/features/gpu/

For Docker Engine, installation of the nVidia container toolkit is required, see https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html