Large scale image processing and high performance computing

Dr Steffen Bollmann
Research Fellow, School of Information Technology and Electrical Engineering and
Adjunct Fellow, Centre for Advanced Imaging, presents: 

Large scale image processing and high performance computing

Researchers require a diverse collection of tools to analyse data and answer their research questions, but often struggle tremendously to run their analyses on adequate hardware and it is very challenging to move analyses between computing platforms due to the setup work required - ultimately limiting interoperability, reproducibility and scalability.

In this talk, I would like to discuss new developments that help to overcome these problems. We developed a modular analysis ecosystem and applied it first in the field of neuroimaging. The first module consists of a continuous integration system using github actions to automatically build various software containers. Our second module provides wrapper scripts that transparently integrate these containers into any existing workflow. The third layer downloads the containers on demand and manages them to keep the system lightweight, and the final layer provides a lightweight Linux desktop accessible via a browser that runs on any operating system.

By this, our project (https://github.com/NeuroDesk/) enables researchers to use any software with any dependency on any operating system. Moreover, as it does not require any installation or specific hardware, it lowers the barriers for entry for less-technical users, while enabling to scale the analysis at later stages.

We developed a scalable and interoperable way of running various scientific tools on any compute platform accessible to researchers, such as our CAI workstations, Nectar, the Characterisation Virtual Laboratory (CVL) or High Performance Clusters.

Register for this webinar

Watch the recorded session in our Seminar Recordings or you can access the presentation slides here.