Accepted to ARRS 2026
Authors: Derek Jiu, Nishant Chinta, Ryan Bui
This study presents a scalable noise injection framework to evaluate how radiographic noise affects chest X-ray analysis. We systematically inject various types and levels of noise into chest X-ray images to measure the impact on semantic segmentation of anatomical structures and disease classification accuracy. Our framework enables researchers to assess model robustness and develop more noise-resilient medical imaging algorithms. [arXiv link TBA]
