SwinFPN: Leveraging Vision Transformers for 3D Organs-At-Risk Detection
Bastian Wittmann, Suprosanna Shit, Fernando Navarro, Jan C. Peeken, Stephanie E. Combs, Bjoern Menze
Current state-of-the-art detection algorithms operating on 2D natural images utilize the relation modeling capability of vision transformers to increase detection performance. However, the feasibility of adapting vision transformers for the 3D medical object detection task remains largely unexplored. To this end, we attempt to leverage vision transformers for organs-at-risk detection and propose a novel feature extraction backbone, dubbed SwinFPN, which exploits the concept of shifted window-based self-attention. We combine SwinFPN with Retina U-Net's head networks and report superior detection performances. Code for SwinFPN will be available in our medical vision transformer library https://github.com/bwittmann/transoar.
Wednesday 6th July
Poster Session 1.1 - onsite 15:20 - 16:20, virtual 11:00 - 12:00 (UTC+2)