Creating a high-quality dataset is a crucial part of any machine learning project. This is even more the case for highly specialized domains such as medicine. In practice, data annotation often takes longer than the actual training and optimization of the models. Medical data annotation includes several process steps: planning, anonymization, provision of the data, annotation/segmentation, and quality assurance. Annotation Lab Essen, an organizational unit of the Institute for AI in Medicine and part of Essen University Hospital, benefits from the highest medical expertise and specializes in the annotation of medical data (images, text and beyond).
Evaluation and Reader Studies of AI algorithms (e.g. on bone age estimation)
Semi-automated pipeline for medical text anonymization for consecutive text annotation
Semi-automatic annotation of histopathological images
Segmentation of pulmonary pathologies of COVID-19 patients' image
Pipeline for generation of structured reports from unstructured radiological data
SOP of Body-Composition Analysis and integration of ontology definitions