In dem regelmäßig stattfindenden Oberseminar tragen Gäste aus aller Welt über Forschungsarbeiten zu Themen vor, die mit der Arbeit von CeVis und MeVis in Verbindung stehen, und Mitarbeiter von CeVis und MeVis präsentieren ihre neusten Ergebnisse.
Detectability of lesions (abnormalities) in medical images is limited by image noise and the complexity of normal patient structure. Some structures can be modeled as correlated noise. In such cases, the upper limit to detection accuracy is determined by the Bayesian ideal observer. This signal detection theory approach has been successfully applied to evaluation of human observer performance of a variety of tasks: detection, discrimination, localization and identification. We are able to perform as suboptimal Bayesian observers with an efficiency of about 50%. That is, we require about 40% higher signal to noise ratio than the ideal observer for a given decision accuracy. This lecture will briefly describe signal detection theory and summarize human experimental results and modeling.