Processing Signals with SNN

Hello everyone, I’m back on the forum again to better understand SNNs.

I am studying better methods of estimating myocardial infarction using SNNs when compared to classical neural networks. My idea is to include processing a layer using SNNs to help with classification, but I’m not having success with my method.

Basically, as in myocardial infarction the ST segment of the heart wave is affected, my idea was to inhibit parts that do not contribute to the classification (for example QRS complex) and maintain a segment of interest that contributes (segment ST). I’m using SNN (with inhibitory neurons) to process the signal and remove the non-contributing parts. I think this method will better classify cases of myocardial infarction, but I have had no success with this approach.

The image below represents the behavior affected by myocardial infarction.

Can anyone help me understand this question, I believe the processing with SNNs is correct and I used Nengo for this case. My biggest doubt is whether really inhibiting non-contributors, if they collaborated with the classification model, in my way of thinking it makes a lot of sense and would collaborate. Could someone who has experience help me understand if it’s worth the effort?

Best regards.

1 Like

Hello,I have the same question as you,have you resolve this question?

1 Like

Hi @RENZHU , I didn’t follow this idea because I don’t have an answer to this problem. I only know how to use this technique to process cardiac signals (i.e. inhibit the QRS complex, T-wave), but I don’t know if it contributes to classification systems using machine learning… See the image below for an example of QRS complex inhibition and maintain a segment ST:

Thank u (: