AI-based COVID test achieves 98% accuracy
Researchers have developed an artificial intelligence algorithm that can detect COVID-19 infections with 98% accuracy using chest X-rays.

Named Custom-CNN, this algorithm provides faster and more accurate results compared to the PCR test. This deep learning-based AI algorithm automatically analyzes chest X-rays to swiftly and accurately identify COVID-19 infections, differentiating them from pneumonia, which often presents similar symptoms.
According to NTV, the PCR test is the most commonly used method for diagnosing COVID-19. However, it has limitations: it's costly, results can be slow, and it tends to produce false negatives.
AI Assists in Simplifying Diagnosis Process
Manually examining X-rays for infection signs is time-consuming and not always accurate due to reliance on human interpretation. Therefore, researchers at the University of Technology Sydney (UTS) turned to AI to facilitate the diagnostic process.
98% Accuracy in Detecting Covid-19
The study published in the journal Scientific Reports revealed that the Custom-CNN model achieved a classification accuracy of 98.19% in distinguishing COVID normal, and pneumonia images.
Early Diagnosis Prevents Further Infections
Early detection of COVID-19 infection can ensure patients receive appropriate treatment, including antivirals most effective within five days of symptom onset.
It also encourages patients to isolate, preventing the spread of infection.
Difficulty in Differentiating Symptoms
Another complication is the challenge of distinguishing COVID-19 symptoms – like fever, cough, breathing difficulty, and sore throat – from other respiratory viral infections such as the flu or pneumonia.
In recent years, machine learning algorithms have gained popularity in medicine, aiding doctors in diagnosing Parkinson's disease, detecting breast cancer, and predicting strokes and heart failure.
Most Read News
-
Iran executes man convicted of spying for Israel
-
Medvedev warns no-fly zone above Ukraine would mean war
-
US sanctions financial network supporting Iran's militar
-
Top US Democrats slam Republican spending bill, warn of
-
Xi, Trump likely to attend Asia-Pacific Economic Coopera
-
Poland's premier blames Russia for house damage amid rep
-
2 British politicians ‘denied entry’ into Israel during
-
EU proposes sanctions on Israeli Cabinet ministers, viol
-
US House committee releases new Epstein documents
-
At least 64 Palestinians killed in fresh Israeli strikes