k-Wave

1. Anthony
   

   Member
   Joined: Apr '13
   Posts: 96

   Hi Brad,

   In the example "Simulating B-mode Ultrasound Images", you model the heterogeneity of the medium by adding scatterers as a (bounded) gaussian noise on speed of sound and density.

   I made a quick bibliographic search and did not find anything precise about modeling the heterogeneities in the tissue. Actually, most of the papers I found are focused on the size (and shape) of the scatterers, estimated from image processing (or histology) but I could not find anything about the statistic distribution of the density and speed of sound among the tissue.

   Do you have some bibliographic references about this or was your example simply based on the visual aspect of the simulated B-mode images, which look realistic?

   I just want to have a look at the defocusing that occurs due to heterogeneity and seems not to be negligible for HIFU simulations.

   Best regards,
   Anthony

   Posted 10 years ago #

2. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,640

   Hi Anthony,

   The values used for this particular simulation were chosen arbitrarily (but hopefully in a vaguely plausible way). I think what is probably most important for imaging simulations is that the backscatter has the right statistics. For example, you could establish a correlation between the standard deviation of the Gaussian noise added to the sound speed/density, and the received backscatter (Szabo's book has some expressions for backscatter if you need them). Then, you could choose the noise statistics for different tissue types based on experimentally measured values of backscatter from the literature.

   One of my students did some work on creating "equivalent scattering distributions" to give the correct backscatter statistics, but we haven't published anything on it yet. Depending on which parameters you are interested in, the scatterer distribution is not unique. For HIFU, you may also have to consider other aspects of the scattering. Unfortunately, I haven't found much literature on defining small-scale tissue heterogeneities for grid-based models.

   Hope that helps,

   Brad.

   Posted 10 years ago #

3. Anthony
   

   Member
   Joined: Apr '13
   Posts: 96

   Thanks for your answer Bradley,

   I will try to play with the parameters and see how the backscatter behaves... A sensitivity analysis on my HIFU field will probably also be a good start.

   Anthony

   Posted 10 years ago #

4. Farnaz
   

   Member
   Joined: Oct '18
   Posts: 20

   Hi,

   Have you published anything on this matter? I have difficulty in setting realistic speckle noise for real tissue (e.g. breast tissue)

   If I were to add uniformly distributed random speckle noise in the density domain with mean distribution of a specific number of reflectors per lambda or lambda^2 (2,3,etc.)(having a 5e6Mhz frequency and speed of sound varying 1300-1800 and dx=3.6765e-5).
   How this would be possible having a grid size not matched with the lambda^2 requirements? Can one create an arbitrary density domain independent of dx?

   Best,
   Farnaz

   Posted 4 years ago #

5. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,640

   Hi Farnaz,

   We haven't done any more work on this. This paper by Gianmarco Pinton might be helpful.

   Brad.

   Posted 4 years ago #

RSS feed for this topic

Reply

You must log in to post.