• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 102 subscribers
  • Views 284 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

[STK-MED]: Regarding Wall filter coefficient generator.

Honey Sukesan
Intellectual 680 points
Other Parts Discussed in Thread: MIDAS

Hi,

I was trying to run custom input data linear and sector with MIDAS. For this, I am analyzing and understanding STK-med code.

I was trying to generate the wall filter coefficient file used in TI MIDAS supplied data sets for heart and carotid using the file MedicalSoftwareToolkitVer2_0_0_usound\usound\sw\dpu\test\vectors\wf_coeff.m.

The generated wall filter coefficients matches with TI_Heart_WallFilterCoeff.bin. So I changed 'N(number of ensembles)" to 10 to test for TI carotid dataset.

But the generated wall filter coefficients don't match with TI_Carotid_WallFilterCoeff_001.bin which is the wall filter coefficient file for Carotid dataset.

Is there anything else to be changed on wf_coeff.m to generate the wall filter coefficients for Carotid? Can I reuse the same wall filter coefficient generator file for just running my custom data with MIDAS?

Thanks,

Honey S

 

  • 0
    TI__Expert 4920 points

    Hi Honey

    I will look into thise and get back to you shortly.

    -- Uday

     

  • 0 in reply to Uday
    TI__Expert 4920 points

    Honey,

    The wf_coeff.m, matlab file that generates wall filter coefficients, includes a call to 'cheby2 (N, R, Wn, type),'  a function from MATLAB's signal processing toolbox, used to design an Nth order filter, where Wn is the cutoff frequency and lies between 0 and 1. See http://www.ece.gatech.edu/research/DSP/courses/ee2200/matlab/10.signal/cheby2.html for more on cheby2.

    The call to cheby2 in wf_coeff.m is as "[b,a] = cheby2(3,90,0.01,'high')" which defines a 3rd order high-pass filter with cut-off frequency Wn as 0.01. I believe we used these parameters to generate the wall filter for heart data, and for the Carotid artery dataset we used a different set of parameters to design the filter (specifically I think we tweaked the cut-off frequency).

    As "An Atlas of Ultrasound Color Flow Imaging, by Goldberg, Merton, Deane" points out, the wall filter setting is variable and is chosen based on what is being imaged. "The wall filter determines the minimum level (mean frequency shift) for color flow information to be displayed by eliminating lower-frequency shifts. The wall filter extends from the baseline to a selected frequency level in both directions, and is often displayed as a black portion on the color dynamic range scale bar equivalent to the selected wall filter level. For venous imaging, the wall filter should be set at a low level to provide greater sensitivity to slow blood flow. High wall filter levels may help to reduce artifacts related to vessel wall motion..which are typically low-frequency and high-amplitude signals." 

    I think wall filter design in itself is a topic of research and differentiation and end-product companies have their own ways of implementing it. As mentioned earlier, MIDAS is an example of a system-level integration and we provide this just as a reference so you can leverage the code to get started quickly on software development, but domain-specific knowledge and design are really left to customers to research and implement.

    -- Uday

  • 0 in reply to Uday
    Prodigy 210 points

    Hi,

    Could you kindly tell me the sampling rate used for calculating the normalized cutoff frequency Wn.

    Regards,

    Mithun