Figure 1 has made a name for itself as a social network that lets medical professionals discuss photos of patient conditions with colleagues around the world.

“They can learn in real time from other people experiencing and seeing cases,” says Dr. Joshua Landy, a practicing physician and cofounder of Figure 1. “If you’re seeing a case, you can take a picture of it, you can describe it and ask for help, and you can even page a specialist.”

Now, the Toronto company sometimes referred to as the “Instagram for doctors” plans to introduce artificial intelligence into the mix, starting with a feature to turn photos of electrocardiograms into digital data. The company is planning to formally announce the feature later this month at the International Congress on Electrocardiology in Portland. At first, experts will be able to weigh in on the meaning of the measurements, but in the future more advanced machine learning systems may be able to provide their own insight into what particular readings mean.

Electrocardiograms translate electrical impulses in the heart into line graphs that doctors can read to diagnose patients. While they’re naturally useful in checking for heart attacks and other cardiac issues, experts can sometimes also spot other conditions in ECG readings, from pneumonia to Parkinson’s disease, Landy explains. And the readings, often considered a vital sign on par with body temperature, blood pressure, and pulse rate, are standardized enough to be a natural target for digital processing.

“They are almost perfect fodder for computer vision and machine learning,” Landy says. “They’re self-similar, they’re stereotypic, they’re immediately recognizable by an algorithm.” full article