The University of Eastern Finland (UEF) has started to organise get-together events around the topics of AI in inverse problems, imaging, and health in the Kuopio region. These events are meant to become low-threshold, community-driven, bottom-up events where students and researchers from different organisations/research groups can meet and share their experiences and get peer-support related to AI-based method development. These meetings are organized under the umbrella of FAME Flagship and DREAM doctoral education pilot, but are open to all interested researchers/students in the Kuopio region (also in companies).
The second meeting of this semester will have a focus on applied AI in medical imaging. We will have two presentations, and time for discussion is reserved after each presentation as that is the main point of these meetings. Coffee will be served!
Programme:
14:00 – 15:00 Aleksei Tiulpin, University of Oulu, “Medical Image Segmentation: The Next Steps”
Abstract: Image segmentation has reshaped medical image analysis, driving major advances in how automatic contouring improves outcomes across domains. Most current methods approach medical image segmentation as a pixel-wise classification task into foreground and background. Yet, what is often overlooked is how uncertainty affects not only model performance but also the reliability of biomarkers derived from these segmentations. In this talk, I will share recent advances from my group, along with joint work with collaborators, highlighting how uncertainty can be harnessed during training and in the estimation of clinically relevant quantities.
15:00 – 15:30 Sheyda Bahnemiri, Planmeca, “Inverse Problems in Medical Imaging: Industrial Applications of AI-Based Scatter Correction in CBCT” (remotely)
