MIDL Vision 2022

The method used here is simple, versatile, and universal . It is not dependent on any particular network architecture or optimization settings. Could this be what swayed the program chairs to pick it for a coveted oral session this year? “ Yes, and our evaluation was quite comprehensive, ” Alexander points out. “ We compared with a broad range of state-of-the-art methods and could demonstrate the efficacy of our simple, elegant method, which is superior to general-purpose domain adaptation methods and methods developed specifically for the pose estimation problem. ” Finally, Alexander tells us how much he has enjoyed working with Mattias on this project. “ Mattias is enthusiastic about everything to do with deep learning and medical applications, ” he reveals. “ He is still writing first-author papers, which is extraordinary for a full-time professor! ” To learn more about this work, come to Oral Session 2.1: Domain Adaptation and Model Generalization today at 09:40 - 10:40 and Poster Session 2.1, onsite at 15:20 - 16:20 and virtual at 11:00 - 12:00 (UTC+2). 14 Oral Presentation VISION MIDL