Al-assisted Imaging Technologies (AIT)

Prof. Dr. Venera Weinhardt [Contact]

Welcome

AI is transforming imaging technologies, particularly in techniques such as X-ray tomography, by accelerating analysis pipelines through deep learning to meet rising demands for high-throughput, large-volume 3D data. In medical X-ray imaging, focused on human medicine, segmentation tools are firmly established; in domains such as uCT and soft X-ray tomography, image segmentation remains a significant bottleneck.

Subject areas

We develop automatic machine learning and deep learning approaches that support the development of X-ray imaging methods.

Automatic segmentation of cell anatomy

In collaboration with Heidelberg's Engineering Mathematics and Computing Lab [link], we develop automatic segmentation tools for 3D segmentation and analysis of anatomy in model organisms and single cells. Current work focuses on extending automatic segmentation of cell and nuclear structures to multi-label networks for complex textures, and on organelle-specific methods leveraging X-ray absorption signatures, shape priors, and diffusion models.

You can find our cytoplasm segmentation model, ACSeg, ready-to-use on your SXTdata at Biomedisa [link]