SDSL: Spectral Distance Scaling Loss Pretraining SwinUNETR for 3D Medical Image Segmentation

Recent approaches utilizing self-supervised learning with masked image modeling (MIM) have demonstrated great performance. However, applying MIM naively to small datasets results in poor generalization to downstream tasks. We hypothesize that capturing detailed anatomical structures can compensate f...

Full description

Bibliographic Details

Published in:	IEEE Access
Main Authors:	Jin Lee, Dang Thanh Vu, Gwanghyun Yu, Jinsul Kim, Kunyung Kim, Jinyoung Kim
Format:	Article
Language:	English
Published:	IEEE 2024-01-01
Subjects:	Spectral bias self-supervised learning swin U-Net transformer medical image segmentation
Online Access:	https://ieeexplore.ieee.org/document/10654210/