Samsung Galaxy Watch7 could use artificial intelligence to detect a major health issue


Samsung Galaxy Watch 6 Pro hero

It seems as if Samsung wants to make the upcoming Galaxy Watch7 the best health-tracking device on the market, at least if we go by rumors and leaks. A recent report suggested that Samsung may include non-invasive blood glucose monitoring on its Galaxy Watch7. Another leak has surfaced online, revealing that Samsung may be working on another neat feature for the Galaxy Watch.

According to a patent filed by Samsung, the company might include the AFib (Atrial Fibrillation) monitoring feature in its Galaxy Watch7. However, the information on the patent doesn’t give us any certainty that the functionality could be ready in time for the Galaxy Watch7. So, it may arrive along with the Galaxy Watch7, the rumored Galaxy Watch ‘Ultra’, or it could be reserved for a future Galaxy Watch.

The functionality to detect AFib (Atrial Fibrillation) could be a big deal if Samsung manages to pull it off correctly. AFib is a condition where the heart beats too slowly, too fast, or in an irregular way, which is very dangerous.

AFib can happen briefly or could be a permanent condition in anyone. As per a report by the Centers for Disease Control and Prevention (CDC), more than 454,000 hospitalizations with AFib are reported each year in the United States.

Samsung is reportedly rumored to be using generative AI (artificial intelligence) to detect AFib. According to the patent, AI could help Samsung convert the measured photoplethysmograph (PPG) signals into ECG signals. However, the patent says that the current models of converting PPG-to-ECG algorithms are “data hungry and not robust.”

The patent notes that the current methods of CardioGAN, which is an attention-based generative adversarial network for generating ECG, provide no evaluation of arrhythmias and require a large amount of data for accurate information.

Moreover, it would be interesting to see how Samsung deploys this technology on a wearable device since such complex learning models are resource-intensive, and wearables are highly resource-constrained.


Leave a Reply

Your email address will not be published. Required fields are marked *