New AI Detects 97% of Lung Diseases, Can Tell Pneumonia from COVID
Researchers have developed a breakthrough AI model that detects lung diseases from ultrasound videos with 96.57% accuracy, distinguishing between pneumonia, COVID-19, and other conditions. Created by teams from Charles Darwin University, United International University, and Australian Catholic University, the hybrid model, TD-CNNLSTM-LungNet, combines convolutional neural networks (CNNs) and long short-term memory (LSTM) models to analyze and explain its findings, aiding radiologists in diagnosis.
The model significantly outperforms existing AI tools, reducing false negatives and enhancing clinical transparency with visual explanations like heat maps. Researchers aim to expand its capabilities to diagnose additional conditions and analyze CT scans and X-rays.