Jakob Hofmann

Vision–Language Models (VLMs) for medical imaging study the joint understanding of visual data and natural language to support tasks such as automated report generation, clinical decision support, and image interpretation. In radiology, VLMs have shown strong potential for generating chest X-ray reports by learning associations between imaging patterns and clinical text. My research focuses on developing vision–language modeling approaches for medical imaging, with an emphasis on automated and reliable chest X-ray report generation.