Using a Fast Fourier Transform

[Chaitu662]

Hi

I'm working on a project which takes up ECG signals and tries to evaluate the condition of the patient.

For one particular disease (ventricular tachycardia) the ECG looks close to a sine wave. Hence, I find the predominant frequency in the signal. I shift the original signal now by half the time period (calculated from the predominant frequency).

Adding the new and the old signal, must give me a value close to zero for this condition only.

Problem is, I'm using FFT (512 samples) to determine the predominant frequency component. Is there any better way to find this since FFT eats up a lot of computing time..

Thanks!

http://en.wikipedia.org/wiki/File:Lead_II_rhythm_ventricular_tachycardia_Vtach_VT.JPG

^Link for ventricular tachycardia.

 

[jvicens]

Hello,

Thank you for sharing your project with us. It sounds like an interesting and important area of research.

Regarding your question about finding the predominant frequency component in ECG signals, there are a few different approaches you could consider. One option is to use a time-domain analysis method, such as autocorrelation or the Hilbert transform, to identify the dominant frequency. Another approach is to use a frequency-domain analysis method that may be more efficient than the FFT, such as the fast Hartley transform or the chirp Z-transform.

Additionally, you may want to consider using a machine learning or artificial intelligence approach to train a model to identify the predominant frequency in ECG signals. This could potentially save time and computing resources in the long run.

I would also recommend consulting with a cardiologist or other medical expert to validate your method and ensure that it accurately reflects the presence of ventricular tachycardia.

Best of luck with your project.