AI system uses geometric normalization for accurate pneumonia and COVID-19 detection in X-rays

Researchers have developed an automated system for identifying pneumonia and COVID-19 in chest X-rays by geometrically normalizing the lung region. The system employs a ResNet-18 model for landmark detection, followed by a geometric normalization process using Generalized Procrustes Analysis and affine warping. A separate ResNet-18 classifier then categorizes images as COVID-19, viral pneumonia, or normal. This approach achieved high accuracy on the COVID-19 Radiography Database, demonstrating that anatomical alignment can yield more robust and artifact-resistant disease recognition compared to raw or artifact-masked images. AI

IMPACT This research demonstrates a novel application of AI in medical imaging, potentially improving diagnostic accuracy and artifact resistance for pulmonary disease detection.

RANK_REASON The cluster contains a research paper detailing a novel method for medical image analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• COVID-19
• COVID-19 Radiography Database
• Delaunay triangulation
• Generalized Procrustes analysis
• Grad-CAM++
• Kermany dataset
• ResNet-18
• Salvador E. Ayala-Raggi

• paper
• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →