peak detect

[tmf171]

Luanna
To answer your question, I have never heard of an ECG until I started reading on the National Instruments Forums, BUT I do a lot of signal processing for Ultrasonic Testing. And well, you have used some of the same mathematics I use, the methods only change slightly.
If you could please give me some data to look at an an explanation of exactly what you want done, not only will it aid me in figuring out the problem and possibly coming to a solution, but it will help you gather your thoughts, and understand your problem a little bit better. To find peaks I have done things such as having the computer find the threshold values and using the array max to pull both the maximum and index from the data. If you do this in a loop it will pull out all the peaks in the signal with a two dimensional array of x and y values. These can then be processed and plotted, etc.

Hope that gets you on your way
Tom

[rpursley8]

Luanna,

The whole point of the peak detect algorithm is to enhance the R wave relative to the rest of the ECG signal.  Granted the peak may be in a different place, but it is different the same way for every ECG cycle.  I wrote this algorithm based on a published paper and have used it successfully many times.  The peaks you see in the enhanced waveform correspond to the steepest slope of the R wave.  The derivative part is used to help distinguish the R and T waves (the slope of the T wave is less than the R wave).  The hilbert transform is used to get an envelope of the signal and therefore a smooth looking signal to peak detect on.

Please post a sample of your data you are having problems with and I will try to help figure out what is going on.

[Luanna]

Hi, Randall and Tom! Here it goes some data, a screenshot comparing the original ecg signal (below) and the enhanced ecg at a specific period, and a screenshot of what i mean by different shifts. I hope it helps. Anything you need me to clarify i´ll be happy to answer.

Thank you for your kind support.
Regards,

Luanna