feat(ecg-app): high-pass filter and pulse detection

the high-pass filter (simply subtracting a moving exponential average) improves the clarity of the graph significantly by eliminating much of the muscle movement etc. that is visible otherwise.

detect_pulse uses this cleaner signal to detect the Q and R spike of the QRS complex to determine the pulse of the user

closed

added 3 - Preload 6 - Needs Review labels

changed milestone to %v1.10

changed the description

I have fixed the lint problems with code-style.sh but cannot find a way to have this merge request retested. Or was there another problem with it @rahix ?

reopened

Ouch, I did not mean to close this, sorry!

assigned to @rahix

Waiting for decision on fuchsi/firmware!1 (merged).

added 2 commits

• 66497153 - style(preload): (ECG) Semantic variable names in pulse detection
• 4f4851ba - Merge branch 'ecg-bpm-counter' into 'master'

Compare with previous version

Looks good to me! Thanks for keeping our codebase clean.

However, I'm only a bot - so a human will still have to approve this request.

Just tried it out, that's really cool!

thanks =) From my side this is ready to be merged. I'm thinking about further improvements for the ecg app (a menu to toggle the bias, the leds, the high-pass filter etc.) but would create a new merge request for those once I start working on them.

(of course if there is something you think should be changed before merging just let me know =) )

Sounds good! Let's wait with the merge for one more person to test it, but otherwise it is looking good!

added 6 - Needs Testing label and removed 6 - Needs Review label

That's really cool, just tested it. I would suggest putting the smoothing & bpm counting into a pycardium module though, as this functionality is very useful for other applications wanting to use the ECG.

I would suggest putting the smoothing & bpm counting into a pycardium module

I think the proper way would be a module in Epicardium which would allow l0dables to make use of it as well. But this is really out of scope for now.

hmm. The high-pass filter I used is certainly not the one way to clean up the signal so I would be hesitant to add it to epicardium. I played around with the parameters until my graph looked as nice as possible (for me) - but even just plotting the differences between samples instead of the samples themselves (so just subtracting the previous value every time) is pretty good already if you just want to see the pulses. So that is probably what I would suggest to people asking me how to clean up the signal. (And it is simple enough that we don't have to offer it in epicardium imho.)

I agree, that offering the BPM directly would be very nice though. I wanted to have a more in-depth look at epicardium anyways but I'm not quite sure what the API for that should look like tbh.

added 6 - To Be Merged label and removed 6 - Needs Testing label

merged