Improving Image Representations via MoCo Pre-training for Multimodal CXR Classification
Improving Image Representations via MoCo Pre-training for Multimodal CXR Classification
F. Serra,Grzegorz Jacenków,2 저자,Alison Q. O'Neil
2022 · DOI: 10.1007/978-3-031-12053-4_46
Annual Conference on Medical Image Understanding and Analysis · 3회 인용
TLDR
It is demonstrated that self-supervised Momentum Contrast (MoCo) pre-training of the image representation model on a large set of unlabelled CXR images improves multimodal performance compared to supervised ImageNet pre-training.