Unpaired Translation of Chest X-ray Images for Lung Opacity Diagnosis via Adaptive Activation Masks and Cross-Domain Alignment

2025-03-25

AI Summary

Pulmonary opacities in chest X-rays (CXRs) obscure anatomical structures, severely degrading lung segmentation accuracy and lesion localization reliability. To address this, we propose an unpaired image translation framework that synthesizes semantically consistent “clear” CXRs from opacity-contaminated inputs. Our key contributions are: (1) an adaptive activation mask that precisely localizes and edits opaque regions; and (2) a cross-domain alignment mechanism enforcing consistency between translated images and the feature space and predicted labels of a pre-trained pathology classifier, thereby enhancing interpretability. Evaluated on multi-center datasets—RSNA, MIMIC-CXR-JPG, and JSRT—our method achieves a Fréchet Inception Distance (FID) of 67.18 (a 68% reduction), a Kernel Inception Distance (KID) of 0.01604 (a 93% reduction), a lung segmentation mean Intersection-over-Union (mIoU) of 91.08%, and lesion detection sensitivity of 97.62%, significantly outperforming state-of-the-art approaches.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Transparent, Interpretable, Explainable ML
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Responsible Web: Algorithmic accountability and transparency on the web
• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Chest X-ray radiographs (CXRs) play a pivotal role in diagnosing and monitoring cardiopulmonary diseases. However, lung opac- ities in CXRs frequently obscure anatomical structures, impeding clear identification of lung borders and complicating the localization of pathology. This challenge significantly hampers segmentation accuracy and precise lesion identification, which are crucial for diagnosis. To tackle these issues, our study proposes an unpaired CXR translation framework that converts CXRs with lung opacities into counterparts without lung opacities while preserving semantic features. Central to our approach is the use of adaptive activation masks to selectively modify opacity regions in lung CXRs. Cross-domain alignment ensures translated CXRs without opacity issues align with feature maps and prediction labels from a pre-trained CXR lesion classifier, facilitating the interpretability of the translation process. We validate our method using RSNA, MIMIC-CXR-JPG and JSRT datasets, demonstrating superior translation quality through lower Frechet Inception Distance (FID) and Kernel Inception Distance (KID) scores compared to existing meth- ods (FID: 67.18 vs. 210.4, KID: 0.01604 vs. 0.225). Evaluation on RSNA opacity, MIMIC acute respiratory distress syndrome (ARDS) patient CXRs and JSRT CXRs show our method enhances segmentation accuracy of lung borders and improves lesion classification, further underscoring its potential in clinical settings (RSNA: mIoU: 76.58% vs. 62.58%, Sensitivity: 85.58% vs. 77.03%; MIMIC ARDS: mIoU: 86.20% vs. 72.07%, Sensitivity: 92.68% vs. 86.85%; JSRT: mIoU: 91.08% vs. 85.6%, Sensitivity: 97.62% vs. 95.04%). Our approach advances CXR imaging analysis, especially in investigating segmentation impacts through image translation techniques.

Problem

• Convert CXR images with opacities to clear ones for better diagnosis
• Improve lung border segmentation and lesion identification accuracy
• Enhance interpretability of CXR translation using adaptive masks and alignment

Innovation

• Unpaired CXR translation framework for opacity removal
• Adaptive activation masks selectively modify opacity regions
• Cross-domain alignment ensures semantic feature preservation