# Gray Level Dependence Matrix applied to time domain signals

It make sense to calculate gray level dependend matrix on al time domain signal in order to extract features related to that matrix?

I have EEG signals from stroke people in the time domain. My goal is to classify from the signal the level of recovery from the stroke (I have also some scale to quantify the recovery).

I tried different classification pipelines but I achieved good results only extracting features related to the GLDM calculated directly from the time courses.

Since I'm no expert in texture analysis I was wondering if make sense to compute the GLDM matrix from a time varying signal instead an image.

I also tried to transform the time signal into a recurrence plot and then compute the GLDM but I didn't achieve good results.

• Hi! Welcome to this site. I'm sorry, I don't fully understand your question. Generally "does it make sense to…" questions are only useful if you define your use case in a little more detail. So if you can, please edit your question to give us a little more context, on what you're trying to achieve, what your signal is, if you have a mathematical model for it, that would be cool to have and so on :) – Marcus Müller Apr 22 '19 at 8:32
• Hi again and welcome to this site. If you received god results you should probably cure cancer. Just kidding, I really mean that you could be clearer. Work on the English level for starters. Explain with mathematical elements what is the exact problem, what did you do, and what is the exact question. Take a look at other questions for reference. We will be more than happy to help if we can but make it easier for us to help you. – havakok Apr 22 '19 at 10:01
• Ah, the English, albeit not completely free from typos, was pretty good (tried to help with that a little bit), @PierpaoloCroce! Soooo, what's a GLDM, and how is it computed? We might be from different areas of signal processing, but as signal processing experts, we're always happy if someone points us to the math, and often can help with but superficial domain knowledge! The problem you're solving and what you seem to have solved so far sound interesting and promising, so this might lead somewhere :)! – Marcus Müller Apr 22 '19 at 14:12