How to Learn from Unlabeled Volume Data: Self-supervised 3D Context Feature Learning

Maximilian Blendowski*, Hannes Nickisch, Mattias P. Heinrich

*Corresponding author for this work


The vast majority of 3D medical images lacks detailed image-based expert annotations. The ongoing advances of deep convolutional neural networks clearly demonstrate the benefit of supervised learning to successfully extract relevant anatomical information and aid image-based analysis and interventions, but it heavily relies on labeled data. Self-supervised learning, that requires no expert labels, provides an appealing way to discover data-inherent patterns and leverage anatomical information freely available from medical images themselves. In this work, we propose a new approach to train effective convolutional feature extractors based on a new concept of image-intrinsic spatial offset relations with an auxiliary heatmap regression loss. The learned features successfully capture semantic, anatomical information and enable state-of-the-art accuracy for a k-NN based one-shot segmentation task without any subsequent fine-tuning.

Original languageEnglish
Title of host publicationMICCAI 2019: Medical Image Computing and Computer Assisted Intervention – MICCAI 2019
EditorsDinggang Shen, Tianming Liu, Terry M. Peters, Lawrence H. Staib, Caroline Essert, Sean Zhou, Pew-Thian Yap, Ali Khan
Number of pages9
Volume11769 LNCS
PublisherSpringer, Cham
Publication date10.10.2019
ISBN (Print)978-3-030-32225-0
ISBN (Electronic)978-3-030-32226-7
Publication statusPublished - 10.10.2019
Event22nd International Conference on Medical Image Computing and Computer-Assisted Intervention - Shenzhen, China
Duration: 13.10.201917.10.2019
Conference number: 232939


Dive into the research topics of 'How to Learn from Unlabeled Volume Data: Self-supervised 3D Context Feature Learning'. Together they form a unique fingerprint.

Cite this