Answer the question
In order to leave comments, you need to log in
Algorithm for measuring heart rate using the phone's camera?
it’s very interesting how such an application works
itunes.apple.com/us/app/instant-heart-rate-free/id409625068
But I didn’t find any information about this method on the Internet.
Maybe someone has information
PS android
market.android.com/search? q=instantheartrate+azumio
Answer the question
In order to leave comments, you need to log in
It uses light intensity analysis measured by the camera. Record a video, select only the intensity channel with some video editor and "stretch" its dynamic range to the maximum. Get a pulsating picture. By counting the period of pulsations, you can determine the heart rate.
To speed up processing, a very low resolution of images taken from the camera is used, according to which the instantaneous intensity value is calculated, averaged over the entire frame, plus a “floating window” analysis that determines a specific current minimum and maximum level. Further, the interval between the maximum level changes is considered - it will be the heart rate with a fairly decent accuracy.
This is only in brief, perhaps the developers of those applications have applied some more interesting algorithm.
It is almost impossible to remove the noise of heart beats with a microphone (too much external interference). The accelerometer is also unlikely to be able to track movements so accurately.
By the way, such devices were previously made using a light bulb and a photodiode (you can look for diagrams in the Radio magazines), they worked precisely by measuring the interval of change in the intensity of the light flux.
I believe the algorithm is not complicated. Something like the following.
We take the latest data in a few seconds. We fix the average values of the light of the image in time. We find the global minimum and maximum. Next, we determine the frequency of the signal.
I doubt something that the camera can normally determine the pulse. The picture came to mind:
For some reason, the microphone comes to mind first. I think it's easiest to hear a heartbeat.
“The authors of the program recommend turning on the LED backlight during measurements. Or, in its absence, take measurements in good light conditions. The camera catches the outflow of blood when you press the lens, and "sees" the pulsation.
But something I doubt.
Or maybe it's just a trick that the iPhone's accelerometer is next to the camera? It seems to me that readings are taken from the accelerometer and the average value is taken over the bursts. But this is just my choice.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question