Logo der Physikalisch-Technischen Bundesanstalt

Artificial Intelligence in X-ray based Medical Imaging

Kolloquium der Abteilung 8

Medical imaging is fully digital today and generates an increasing amount of high-quality imaging data. To evaluate these data in an efficient manner, artificial intelligence can be integrated at all stages of data processing and analysis, ranging from image formation to analysis to diagnosis prediction and reporting.

This presentation covers a short introduction to artificial intelligence, supervised learning, and physics based forward models as well as examples for deep learning applications in digital X-ray imaging and computed tomography. Image formation examples are focused on noise and dose reduction, artefact reduction, and detection, quantification, and compensation of motion. Applications of deep learning to image analysis, quantification and diagnosis prediction are further areas demonstrated as well as deep learning methods to integrate combinations of imaging and non-imaging data.