Data Science and Beer (online)

Hello everybody,

we look forward to meet all of you for two awesome talks on November 25th at 6 pm:

• "Introduction into Graph Neuronal Networks" (Simon Wein)
• "A 2D slice-based multi-input convolutional neural network for Multiple Sclerosis lesion detection in multimodal brain MRI scans" (Florian Raab)

Zoomlink will be sent 2 hours ahead to all attendees.

See you then!
Severin, Elmar, Michael and Andrea

---
Abstracts:

"Introduction into Graph Neuronal Networks" (Simon Wein)

Inspired by their applications in computer vision, convolutional neural networks were recently generalized to data with graph-like structures. Graph neural networks allow to find patterns in the irregular geometry of graphs, what allows us to apply neural networks to various new learning problems. This presentation gives a short introduction on the concepts of graph neural networks and shows an application for the analysis of brain networks in magnetic resonance imaging.

--

"A 2D slice-based multi-input convolutional neural network for Multiple Sclerosis lesion detection in multimodal brain MRI scans" (Florian Raab)

In this talk, a study to create a fully automated, reliable, robust and fast multiple sclerosis lesion segmentation tool for multimodal MRI scans using deep learning will be presented. The proposed network helps to segment the multiple sclerosis lesions in the brain accurately and can be helpful especially for follow-up scans to monitor the progress of the disease. This network can perform this task in just a few seconds and is a valuable tool for treatment planning (e.g. changes in basic therapy).