Source: ATYUN AI platform
This algorithm allows physicians to more easily predict the risk of cardiovascular disease by using a retinal scan.
Google and its subsidiaries Verily scientists have discovered a new machine learning methods to assess a person's risk of heart disease. By analyzing the results of the scan of the back of the patient's eye, the company's software can accurately infer data, including the individual's age, blood pressure, and whether they smoke. This information can be used to predict their risk of heart disease, such as heart attack, and is currently the main method is substantially the same.
This algorithm may make doctors more quickly and easily analyze the patient's risk of cardiovascular disease because it does not require blood tests. However, this method needs to be tested more thoroughly in clinical applications.
- A paper describing this work: https: //www.nature.com/articles/s41551-018-0195-0
Medical University of Adelaide researcher Luke Oak Dana - Reina think the work is reliable, and shows how artificial intelligence is to help improve the existing diagnostic tools:
"They Data (artificial intelligence) is due to the acquisition of a clinical reason obtained, and than we are currently doing to more than replace the doctor, they are trying to expand what we can do."
In order to train the algorithm, Google and scientists analyzed nearly 300,000 patient's medical data using machine learning techniques. This information includes general medical and eye scan data. Like all depth analysis of the same study, the neural network model was used to dig the information, learn the signs of eye scans and predict cardiovascular risk (such as age and blood pressure) indicators linked.
Although your eyes to judge your heart healthy idea may sound unusual, but it is derived from a study of established. After the inner wall of the eye (fundus) filled reflect the overall health of the body's blood vessels. By using a camera and a microscope to study their appearance, doctors can infer your blood pressure, age, and smoking, which are an important predictor of cardiovascular health.
Two photographs of the fundus. On the left is an ordinary image; on the right shows how Google's algorithm selected blood vessels (green) to predict blood pressure.
When the retinal image is presented in two patients, one of the patients in the next five years, cardiovascular disease, and one patient who was not, Google's algorithm can distinguish correctly in 70% of the time. This scoring method (the need for blood tests and make the correct prediction in 72% of the time) than the commonly used to predict risk of cardiovascular slightly worse.
Alun Hughes cardiovascular physiology and pharmacology professor at University College London, he said, Google's approach sounds very credible, because "in this area, which has long focused on the study of retinal risk prediction of cardiovascular disease use." He He added that AI is possible to speed up the analysis of existing medical form, but he warned that this algorithm requires further testing to be trusted.
For Google, this work is not just a new method of determining cardiovascular risk. It provides a new artificial intelligence-driven paradigm for scientific discovery. While most medical algorithms are used to copy existing diagnostic tools (such as identification skin cancer), but the algorithm to find a new method for analyzing existing medical data. With enough data, people want to be able to create a new artificial intelligence medical opinion without human guidance situation.
For now, doctors churning out artificial intelligence newly diagnosed without human supervision case the idea is a distant prospect, but Google research shows that this idea is not entirely unrealistic.
This switched ATYUN artificial intelligence media platforms, the original link: Observe your eyes will be able to predict heart disease? Google's new artificial intelligence algorithms can be done
more recommendations
H2O.ai: machine learning platform provides critical features for Acordo Certo
NIST released a draft standard AI program in response to Trump's executive order
BloomReach: consumer-focused company will appeal Bloomreach retail investment digitization
