The "Time of High Technologies" program on the YASNAE TV channel aired a segment on how neural networks aid in the early diagnosis of diseases. Vasily Kovalev, a leading researcher at the Laboratory of Biomedical Image Analysis at the Unite Institute of Informatics and Problems of the National Academy of Sciences of Belarus, shared his opinion on the importance of artificial intelligence in modern medicine.

According to him, researchers are currently focusing on three types of images: chest X-rays, computed tomography (CT) scans, and histological images.

"AI methods are very powerful, but one of the major requirements is that training a neural network requires a large amount of input data," the expert noted. He explained that a single image is sufficient for an X-ray, while a CT scan requires analyzing multiple slices—up to 300 images. Each image requires a specialist at least 10 seconds to analyze.

Despite advances in neural networks, Vasily Kovalev emphasized that neural networks are not yet ready to replace specialists. The concept of "computerized diagnosis" is absent from international practice. "There are computerized diagnoses, where AI assists the doctor but does not replace them," he concluded.

Thus, neural networks are becoming an important tool in diagnostics, but medical expertise and experience remain indispensable.