From Code to Clinic
- Colloquium

From Code to Clinic
AI algorithms are becoming indispensable in healthcare. They are on the verge of transitioning from unimodal, single-task applications to versatile, multimodal, general-purpose tools. Developing and deploying such advanced AI companions requires a multidisciplinary approach and a suitable ecosystem. This talk will explore the construction of these systems, covering applications ranging from medical image processing to the use of clinical language foundation models. The importance of interdisciplinary collaboration and real-world implementations will be highlighted with practical examples.



![3D visualisation of human neuronal tissue reconstructed by multi-scale X-ray phase contrast tomography. Neuronal cell nuclei are shown in yellow for the granule neurons in the dentate gyrus region of the hippocampus. Blood vessels are shown in red. By changing the X-ray optical magnification in the multi-scale recordings, one can zoom into regions-of-interest (red ovals). In these scans the resolution is high enough to resolve sub-structures of the nucleus, associated with different DNA packing regimes. Adapted from [6]](/storages/physik/_processed_/e/4/csm_Kolloquium_Salditt_0e30a3f090.png)




