|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- en |
|
|
pipeline_tag: image-to-image |
|
|
tags: |
|
|
- Chest X-ray |
|
|
- Data Augmentation |
|
|
- Counterfactual Image Generation |
|
|
- Bias Mitigation |
|
|
- Medical Imaging |
|
|
- Computer Vision |
|
|
--- |
|
|
|
|
|
# Generative Counterfactual Augmentation (GCA) |
|
|
|
|
|
We explore how to construct unbiased Chest X-Ray datasets using StyleGAN! |
|
|
 |
|
|
<img width="1035" alt="Screenshot 2025-03-29 at 3 05 31 AM" src="https://github.com/user-attachments/assets/3839dcf0-6094-448e-916c-f7635a668ca7" /> |
|
|
## OverView |
|
|
### Introduction |
|
|
This is the pytorch version of previously proposed GCA algorithm. Debiasing large chest X-ray datasets using StyleGAN3. |
|
|
We use code from HiddenInPlainSight repository to demonstrate our methods robustness agains adversarial attacks |
