ecg baseline wandering[need quick expert help]

[Stefan03]

Hello everyone!

I'm new user here, and I have an urgent problem considering my faculty project.

I have looked at previous topics on this matter, and also I've checked the tutorial here

http://zone.ni.com/devzone/cda/tut/p/id/6349

 but I didnt find the exact thing I need.

So here is the thingd:

I have baseline wandering, but how can I fix it without using additional 

 Digital Filter Design Toolkit (DFDT). I have access only to labview 8.5 at laboratory at my faculty

and I dont have DFDT there.

I have attached what I have done so far, and it would mean a lot if someone could help me 

by explaining me how to create/or create for me the filter for wandering.

I would upload the ecg txt file that I use but it's 35mb.

Thanks in advance people!

[Stefan03]

Here is the printscreen of my signal. You can se that peak detection is not working when I have wandering.

[Steve.J]

Hi Stefan,

You don't need the DFDT - there are standard digital implementations of IIR and FIR filters already included in LabVIEW.  They may not be included in the "base" package, but they are standard with the version of LabVIEW that has all of the signal processing functions.  Look in the Signal Processing palette and then in the Filters sub-palette.  Choose a filter like "Butterworth" and configure it as a high-pass filter (Filter Type input) with a cut-off of about 0.5Hz.   Experiment with different filter types, filter order, and cut-off frequency until you are satisfied.  The DFDT helps to speed this experimentation since you can designate certain requirements and watch the results on a Bode plot and the toolkit will build a filter that meets your requirements, but the end results are the filters that are already in this palette.

Steve

[Coq_Rouge]

I agree with Steve. Also remember that if your goal only is to extract heart rate you will only be interested in the QRS complex. The rest of the ECG complex that have lower frequency content may be shaved away with filtering. In your signal you have a large bump just before sample 8000. This is caused by  electrode movement. Securing the electrodes is important.  But your biggest problem is that you do not use the "Peak Detector.vi" correct.

Here is some useful tips for the Labview beginner (we have all been there some time)

1)Then starting a new Labview session enable context help (Ctrl+h) Do not be afraid to use the detailed help option

2)Then stuck, go to help in the toolbar then select find examples

Ps. I also noticed that your Labview programming style is somewhat mixed-up. It is hard to follow the data flow. 

[Stefan03]

First of all thank you very much for such quick responses.

I must admit that I am newbie at this, but I am trying 🙂

So thats why I have furher questions:

-Steve, I have tried Butterworth filter, and all kind of other filters, but I wasnt able to get anything on the waveform graph. 

Do you have any idea why is that? Can you suggest the best way for realisation of wavelet form filter, something that you have come across in your expirience?

-Coq rouge, I know why this errors appeared, but they wanted us to have this kind of signal so we should fix it.

What am I using wrong with peak detetcor?  I have used it on the normalised signal and it was working fine.(attachment)

I though I should only use the baseline correction and than it would be ok.

If you have any suggestions or criticsism that I could use, please post. I know I should use help, but some times time factor doesnt let me.:)

Thanks a bunch guys!

[Steve.J]

Can you attach the updated VI with the filter inserted and we'll take a look at it.

[Stefan03]

Ok no problem, here it is. I just want to filter the segment located lowest in my .vi file(the one with the Butterworth filter).

You dont have to look on the rest of the program.

Here is the link with the txt files of the ecg signals I use:

http://rapidshare.com/files/232956893/EKG_signali_14-05-09.rar

Message Edited by Stefan03 on 05-18-2009 11:59 AM

[Coq_Rouge]

For a starter you have not tied anything to the "sampling freq: fs" input on the filter VI. You can not program labview by tossing in some VIs in a "happy go lucky" style. A very good method for learning Labview is to study examples. For a starter I strongly recommend that you take some time and look at some examples regarding use of digital filters, and the peak detector. Then you may come back and post your questions. You should also find out which sample rate they have used in your sample file

Good luck and have fun in the meantime

[Claire_Reid]

This might be far off, but you might want to look at the Biomedical Startup Kit.