Color Doppler Mapping with Raw Data

[davidjuarezch]

Hello! I am currently working on using the raw data from the device to calculate its doppler frequency, specifically the IQ data, of course. I have the following questions and would greatly appreciate your thoughts and/or guidance:

1. Does it make sense to capture the images with Doppler mode activated or just take B-mode videos and use their IQ data to calculate the velocities and perform the color Doppler mapping?
2. Do you happen to have any guidelines in general on how to correctly process IQ Data and/or calculating the doppler frequencies to calculate blood flow velocity? I've been using a slow time autocorrelation method (over the slow time axis) but have not been successful so far.

In general, any sort of input is helpful, as has been the entire repository up to this point for me!

Thanks in advance, David

[clariusk]

1. you would want to use Doppler mode, as the B mode PRF across frames would be too low to pick up any meaningful signal. The color Doppler data is captured in the set PRF on-screen, with a typical ensemble of 8-12 packets

2. wallfilter (typically IIR) + autocorrelation

[davidjuarezch]

Thanks for your response! I have been working for the past weeks using captured Doppler mode video images as well as using the doppler example in this repository and haven't been able to get an accurate color Doppler map yet. I have, with my own data, been able to get reasonable velocity values. My question is, is there any other processing that must be done to the IQ data before the wall filter and posterior autocorrelation process? I am currently just reading the IQ data with the files provided in the repository and working with it from there.

[clariusk]

typical doppler processing is wall filter, auto-correlation, then thresholding on variance or power based on the signals that you want to extract

[davidjuarezch]

Makes sense. I also wanted to clarify something, which might help with what I'm doing. When using the 'runme' file, why is it not possible to generate a B-Mode image from IQ Data when Color/Power mode is activated, contrary to when it is deactivated? I'm trying to understand it, so I get a grip of how the IQ data is structured to be able to generate color Doppler maps from it, as I have followed typical Doppler processing but still unsuccessful. This is an example of how the IQ Data (Doppler mode activated) looks when I run it through the 'runme' file.

Thanks again for your responses, very much appreciated.

[clariusk]

the IQ data captured from Doppler will not render good B mode since sequencing and parameters are completely different, compared to when you're using the IQ generated for the B-mode sequenced frame