k-Wave

1. eghbalam
   

   Member
   Joined: Mar '15
   Posts: 3

   Dear Dr Cox and Dr.Treeby,
   Hi

   I have a question about Image reconstruction using K-wave toolbox in Matlab and was wondering if you could answer.
   This toolbox has produced an image for me using photoacoustic signals that I have determined as the inputs of the toolbox. These signals are simulated pressue signals (in Pa) detected on the transducers. However, in experimental works, we detect voltage signals (in V) from the transducers. Is it possible to import these voltage signals to K-wave and produce an image of the tissue, and if yes, how can it be done?
   Thank you very much.

   Posted 9 years ago #

2. bencox
   

   Developer
   Joined: Mar '10
   Posts: 292

   Hi eghbalam,

   Are you concerned that you have voltage and not pressure measurements, or can you not work out how to put your recorded time series into the reconstruction algorithm?

   For the former, if you are using a calibrated detector it should come with a known sensitivity in Volts/Pascals so you can convert directly. For the latter, you need to know the positions of the detectors (eg. in Cartesian coordinates relative to some origin) and input them and the recorded time series into kspaceFirstOrder3D, for a time reversal reconstruction (if that's what you're using). One way to do the reconstruction is to follow this example in the Toolbox, and putting your data and sensor positions in the relevant places. Another would be to use kspaceFirstOrder3D with source.p_mode set to dirichlet and feed your measured signals, in time-reversed order, into source points corresponding to your detector locations. You may find the function cart2grid helpful too.

   Good luck with the reconstructions!
   Ben

   Posted 9 years ago #

3. eghbalam
   

   Member
   Joined: Mar '15
   Posts: 3

   Dear Dr bencox,

   Thank you very much for your complete answer. I have already done image reconstruction using simulated pressure signals and completely know how to use these signals to reconstruct PA images. I am concerned about using voltage signals for image reconstruction. As you know, Pressure signals are N-shaped. Unfortunately, I have not done experimental works and have not seen the voltage signals from the transducers. Are the photoacoustic voltages N-shaped, too? Is the only difference between photoacoustic pressure and voltage signals their amplitude?

   Thanks a lot,
   Sincerely,
   Eghbal Amidi

   Posted 9 years ago #

4. bencox
   

   Developer
   Joined: Mar '10
   Posts: 292

   Hi Eghbal,

   The difference between the photoacoustic signal arriving at the detector and the voltage signal measured from the detector can only be due to the response of the detector. If the detector were omni-directional and had a flat frequency response (which is never the case in practice) then the only number you would need to know would be the calibration from Volts to Pascals. In practice, the amplitude of the response at the detector will depend on the frequency and the angle of incidence of the wave. The frequency response is probably the more important of the two to worry about. Photoacoustic signals are inherently broadband - contain a wide range of frequencies - but some detectors are narrowband and record over a restricted frequency range. If you know the frequency response of your detector then you can include that in the model and see what you would expect the time series to look like. Ideally, for photoacoustic measurements, you want as broadband a transducer as you can get.

   Does that help?

   Ben

   Posted 9 years ago #

5. eghbalam
   

   Member
   Joined: Mar '15
   Posts: 3

   Dear Dr. Cox,

   I do not know how to express my thanks to you for your great help. Your information will definitely help me. However, I am still a little bit confused. Suppose that we have observed the voltage PA signal on oscilloscope. In order to reconstruct the PA image using K-wave, do we need to convert this signal into Pressure via frequency response of the transducer? How can it be done? Is it necessary to use mathematical concepts like convolution to do this task?
   Thanks again.

   Sincerely yours,
   Eghbal Amidi

   Posted 9 years ago #

6. bencox
   

   Developer
   Joined: Mar '10
   Posts: 292

   Hi Eghbal,

   As a first step, you can just scale the voltage signal by dividing it with the known sensitivity of the detector, which will be on the calibration certificate in Volts/Pa. If you know this sensitivity as a function of frequency, you could scale it differently at each frequency. You'll find further discussion of this in books on signal processing.

   Regards,
   Ben

   Posted 9 years ago #

RSS feed for this topic

Reply

You must log in to post.