Detecting COVID-19 from X-ray Images

AI has the potential to revolutionize healthcare by classifying, detecting and segmenting diseases in radiology image data so that doctors can focus more on patient care and treatment. More so recently, the medical system worldwide is overloaded. There are not enough doctors to fulfill the demand, and they are heavily impacted by workflows, bureaucracies, and billing processes. Today, doctors, radiologists and pathologists spend hours upon hours applying annotations to image data, manually, and at a very high cost. Annotation or labeling is the principal obstacle in making AI models, which can be reduced 90% by AI. The secret is that only a few, typically 10%, of the images in a dataset contain virtually all the information needed to label the data, accurately – but finding them is challenging. These annotations are made very quickly by Samsung SDSA’s autoLabel software, which effectively sorts the images in the correct order that provides the most information.

Diagnosing COVID-19 usually requires a nasal swab and a laboratory analysis that takes time. By labeling only 6% of COVID-19 X-ray images, Samsung SDSA’s autoLabel converged label accuracy 94% faster than random order data labeling and 50% faster than regular active learning found in other AI platforms. The resulting COVID-19 classification model achieved 95% accuracy which is 4% more accurate than Rapid Test Kit nasal swab, and the Rapid Test Kits are often not available. Samsung SDSA’s COVID-19 X-ray classifier is also 1% more accurate than the PCR test that takes 1 – 2 days for results, and lung X-rays are quick, easy, and cheap.
Additionally, a doctor’s or radiologist’s opinion might be wrong about 30% of the time, and so complex diseases – diagnosed from complex medical images – are often either missed or misdiagnosed. Getting a second opinion is already a billion-dollar market, today, and second opinions typically come from other humans. Artificial Intelligence provides an objective third party and very neutral second or third opinion to any diagnostic procedure. Using AI removes the variability of any one physician having a certain amount of experience, because the AI model can process vastly more cases and does not forget things. Click to learn more…