k-Wave

1. Pavel
   

   Member
   Joined: Nov '12
   Posts: 7

   Dear Dr. Treeby and Dr. Cox,

   Could you please advise if K-Wave allows to define detectors that overlap with each other in space? For example, after laser shot #1 you measure A-scan by finite-size detector with aperture D. After laser shot #2 you slightly move your detector at distance DX over the tissue,but DX does not exceed the size D of detector (DX<<D). In real world you can not use two detectors and one laser shot to obtain two A-scans simultaneously, since such detectors will physically overlap.

   Does your K-Wave allow to use "single laser shot" for big number of overlaping detectors? If it could allow that, that would drastically reduce computation time in problem that we try to solve now using K-Wave.

   Thank you and Regards,
   Pavel Subochev,PhD

   Posted 11 years ago #

2. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Hi Pavel,

   The sensor mask that you define in k-Wave is just a collection of grid points where the data is recorded during the simulation. One possibility would be to define a sensor mask that includes all the possible detector positions, then process the data afterwards into individual A-scans using overlapping points. For example, if your sensor mask is a line, you could use something like data = sensor_data(start_index:start_index + width, :) and then increment start_index as you wish.

   Brad.

   Posted 11 years ago #

3. Pavel
   

   Member
   Joined: Nov '12
   Posts: 7

   Hi Brad,

   Thank you very much, it really makes sence for the line sensor in your example. However, we would like to define more complicated sensors in 3D. In order to compute all possible sensor points it seems reasonable to define sensor mask as "sensor.mask(:,:,:)=ones(1:Nx, 1:Ny, 1:Nz)", so that we can have 4D array "sensordata4D(1:Nx, 1:Ny, 1:Nz, 1:Nt)" that would contain A-scans in all possible sensor points. Then we can combine them as we need.

   Unfortunately, K-Wave output saves everything in 2D array which combines A-scans from all the sensors as sensordata2D(1:Nx*Ny*Nz,1:Nt).

   Could you please advise if there is sure way to extract the data from 2D array "sensordata2D(1:Nx*Ny*Nz,1:Nt)" into 4D array "sensordata4D(1:Nx, 1:Ny, 1:Nz, 1:Nt)". Is there any mixing up between different elements and their numbers in sensordata2D?

   Would it correctly work as the following:
   for i=1:Nx; for j=1:Ny; for k=1:Nz;
   sensordata4D(i, j, k, 1:Nt)=sensordata2D(i*j*k,1:Nt);
   end; end; end;

   We can use "ijk cycle" and combine sensordata4D from point-like (xi,yj,zk) detectors, but it ruins initial idea to shorten the computational time. Using just "ij cycle" to combine sensordata4D from line detectors still requires too much time. This is all we tried yet.

   Thank you and Regards,
   Pavel Subochev

   Posted 11 years ago #

4. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Hi Pavel,

   If you are storing the complete domain, the output sensor data will be stored in column major order, indexed as x, then y, then z. You could reorder the output data using reshape, for example:

   sensordata4D = reshape(sensordata2D, [Nx, Ny, Nz, Nt]);

   I hope that helps,

   Brad.

   Posted 11 years ago #

RSS feed for this topic

Reply

You must log in to post.