Self-supervised Training in chest X-rays domain.
Principal Investigator
Name
Song Luo
Degrees
M.S
Institution
Northwestern University
Position Title
Research Assistant
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCOI-920
Initial CDAS Request Approval
Feb 16, 2022
Title
Self-supervised Training in chest X-rays domain.
Summary
As Self-supervised Training became popular in deep learning especially associated with computer vision, the desire to collect massive data grows. In this project, we will apply self-supervised pretraining on multiple deep learning models in chest X-rays (CXRs) domain. To this end, we want to collect the publicly available CXR datasets as much as we can. After we have collected the data, we will compare our results to the ones from previous literature.
Aims
1. Collect and combine world's largest CXRs dataset.
2. Apply self-supervised training method on the dataset.
3. Compare results to the previous ones.
Collaborators
Han Liu,
Zhihan Zhou