VoxSeP: semi-positive voxels assist self-supervised 3D medical segmentation

Medical image segmentation enjoys the advantage of understanding 3D contexts, but 3D networks are prone to over-fitting due to the limited amount of annotated data. This paper investigates self-supervised pre-training, i.e. , making use of unlabeled medical data to initialize 3D segmentation network...

Full description

Saved in:
Bibliographic Details
Published in:Multimedia systems 2023-02, Vol.29 (1), p.33-48
Main Authors: Yang, Zijie, Xie, Lingxi, Zhou, Wei, Huo, Xinyue, Wei, Longhui, Lu, Jian, Tian, Qi, Tang, Sheng
Format: Article
Language:English
Subjects:
Computer Communication Networks
Computer Graphics
Computer networks
Computer Science
Cryptology
Data Storage Representation
Datasets
Image segmentation
Medical imaging
Multimedia Information Systems
Operating Systems
Regular Paper
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Medical image segmentation enjoys the advantage of understanding 3D contexts, but 3D networks are prone to over-fitting due to the limited amount of annotated data. This paper investigates self-supervised pre-training, i.e. , making use of unlabeled medical data to initialize 3D segmentation networks. We build our system upon contrastive learning, where the dependence on positive and negative samples obstructs it from satisfying performance on medical image datasets with fewer samples. To alleviate this issue, we present a novel proxy task that takes advantage of the property of human body similarity in medical scans, and defines the sub-volumes from the same position of different cases as Semi-Positive samples. Pre-trained on a mixed dataset containing 1254 CT volumes, the proposed approach, VoxSeP , transfers well to 4 downstream datasets with 2 different backbone networks. On both fully supervised and semi-supervised fine-tuning, VoxSeP achieves favorable averaged improvements ( 2 % and 4 % ), which surpass several state-of-the-art counterparts.
ISSN:0942-4962
1432-1882
DOI:10.1007/s00530-022-00977-9