AI-based COVID test achieves 98% accuracy

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.