Generative model to convert PPG signal to ECG has been tested and works
Samsung Galaxy smartwatches could benefit from continuous monitoring for heart conditions, judging by a new patent filing from the tech giant.
While current devices like the Galaxy Watch 6 enable manual electrocardiogram (ECG) readings to check for irregular heart rhythm – and identify life-threatening conditions like atrial fibrillation (AFib) – this requires the wearer to press a fingertip against the dedicated sensor actively.
However, Samsung has identified a way to translate your regular heart rate readings into ECG readings, enabling that metric to be continuously and automatically tracked.
> Samsung Galaxy Watch 7 rumors and news
The new algorithm is described in a patent spied by Wareable, entitled “System and method for continuous atrial fibrillation detection via PPG to ECG signal translation.”
We’ve seen similar features launched by Fitbit, which can continuously scan for suspected atrial fibrillation via the PPG sensor, and prompt the user to take an ECG. Likewise, Apple also has a similar feature, but limits its use to people who have been diagnosed with Afib.
How it works
Under the plans, the ECG signal would only be constructed rather than actively measured. Still, Samsung reckons it has the qualitative and quantitative evidence to demonstrate the effectiveness and efficiency of the technology.
“In an embodiment, a method is provided for AFib detection,” the patent reads. “The method includes receiving PPG signals of a user measured by a PPG sensor; translating the measured PPG signals into ECG signals using a dynamic model; analysing the translated ECG signals using an AFib detection model, which is trained on measured ECG signals for AFib detection; and providing the analysed AFib detection results to the user.”
Samsung explains that although there is already a strong correlation between the PPG and ECG readings, necessary accuracy doesn’t exist in established methods for a reliable clinical diagnosis. Current attempts to overcome this have proved too cumbersome and inefficient for consumer wearables, the company says.
And it’s good ol’ generative AI that’s here to help. More specifically “a generative model which incorporates prior knowledge about data structures that enables data-efficient learning.”
Samsung explains that “a subject-independent attention-based deep state-space model is provided to translate PPG signals to corresponding ECG waveforms. The model is highly data-efficient by incorporating prior knowledge in terms of probabilistic graphical models. The model may be used to detect AFib, which is one of the most common heart rhythm disorders in adults, by complementing ECG accuracy with continuous PPG monitoring.”
While PPG sensors are currently used for measuring heart rate, blood oxygen saturation and other such metrics within most health trackers, adding a means of accurately simulating an ECG measurement would mean “clinical diagnoses of cardiac diseases and anomalies can be made in real-time.”
Samsung says: “In accordance with an embodiment of the disclosure, the close correlation between ECG and PPG is used to develop an efficient method for synthesizing ECG from PPG waveform, providing an economical and user-friendly ECG screening for continuous and long-term monitoring. This method takes advantage of both the rich clinical knowledge base of ECG in many instances, ranging from well-established ECG-based arrhythmia detection and abnormalities of heartbeat in unhealthy patients, focal causes of atrial fibrillation, and ventricular tachycardias, as well as the convenience and accessibility of PPG in daily settings.”
It appears that Samsung is quite far along with this technology as it has qualified the accuracy and effectiveness of its algorithm within the patent. Whether it is rolled out to the Galaxy Watch series (it’ll also work on other wearables that can currently detect PPG signals – i.e. the forthcoming Galaxy Ring) in a future generation remains to be seen.
Samsung would need FDA approval for the tech for it to be considered a medical-grade device. So, it’s unlikely to replace the ECG sensors already sitting within its smartwatches. However, continuous monitoring would be a great bonus to have and could potentially save lives.