Deep Learning for Chest Radiograph Diagnosis in Africa We have developed a deep learning algorithm to concurrently detect COVID-19 and 14 other clinically important diseases in chest radiographs. We focus on the African continent where only 0-5 CT units are installed per 1,000,000 inhabitants and aim to become the number one AI medical image software company in Africa. Our algorithm, CheXRad, uses 15 machine learning models and is trained to predict diseases on X-Ray and CT images and highlight parts of an image most indicative of each predicted disease. We found that CheXRad can diagnose certain pathologies at a level comparable to practicing radiologists. We have achieved 99.91% AUC (Area Under the Curve) and are working towards 1 million X-Ray and CT scans to further enhance our algorithm. For more accrad.com
Member count: 1-10