[ad_1]
A latest report released revealed that artificial intelligence can at best be equivalent to human experts in medical diagnostics from images. Today, AI research in medicine is advancing rapidly and has the potential to transform medical research from below. Experts say this will alleviate resource constraints, free up time for doctor-patient interactions, and facilitate the establishment of personalized diagnoses.
Given that the sector is overwhelmed by impecunious research, the latest findings are based on a few studies. The researchers discovered that one of the booming applications is the use of artificial intelligence in medical images. This domain is based on Deep Learning, in which a series of tagged images is introduced into algorithms that extract features and learn to classify similar images.
This approach has been promising in the treatment of diseases, eye diseases to cancer. However, in response to questions about how such deep learning systems compare to human skills, the researchers conducted the first comprehensive review of published studies on the subject, revealing that humans and machines were at risk. equality.
According to Professor Alastair Denniston, one of the founders of the Birmingham NHS University Hospitals Foundation and co-author of the study, the results have been encouraging, but this study has helped to verify if any of the hype around the AI was in fashion.
In their writings on the Lancet Digital Health, Denniston, Dr. Xiaoxuan Liu, lead author of the study, and his colleagues pointed out how they focused on research papers published since 2012, an essential year for the ### 39, deep learning. In more than 20,000 German studies, only 14 studies, all based on human illness, reported good quality data, tested the system thoroughly with images from a separate dataset of the one used for training, and presented the same images to human experts.
Then, the team gathered the most promising results from each of the 14 studies that found that deep learning systems correctly detected disease status 87% of the time, compared to 86% for health professionals. % of the time that outperforms the human experts with 91%.
With these results, Denniston is very optimistic about the potential of artificial intelligence in health care, claiming that such deep learning systems could be a diagnostic tool and help reduce the risk of artificial intelligence. backlog of scans and images. In the same context, Liu said that such systems could prove useful in places where there is a lack of experts to interpret the images. It would be important to take advantage of deep learning systems in clinical trials to determine if the results for patients have improved compared to current practices.
Share this article
Do the sharing trick
[ad_2]
Source link