I'm inclined to think that with a proper signal
processing chain (e.g., combination of linear
and rank-order time series filter) one could
do a pretty respectable job of untangling the
baseline shifts from the pulse beats. The
signal characteristics are different enough that
a source-separation relaxation algorithm ought
to also converge fairly well, unless the
subject is pretty active. Something like
walking or running would be troublesome.
Linear filters would be hopeless without
a cooperative subject.
- Doug K
>Not too practical if the person moves even slightly!
>>The deviced described measures light absorption by
>blood. As the heart beats, the volume of blood in the
>light path changes, changing the intensity of transmitted
>light. Movement of the finger within the device kills the
>signal until the electronics can establish a new baseline.
>>The real question is whether you intend to monitor heartrate
>on a cooperative subject or are trying to do it surreptitiously.