Goal for next stage and a progress update

[Jean Rintoul]

Hi all, 

I was really excited to see so many people interested in this project at 34C3, where this project was first announced. 

The initial goal is improved spatial resolution to the point whereby the seeds in an apple can be detected, and the 3D apple shape reconstructed! 

I'd love your feedback on the question at the bottom of this email. 

For improved spatial resolution we'd need:

1. Improved mathematical reconstructions.
2. A new ME repository, which encompasses a range of 3D printable physical set ups, potentially including stepper motors for rotation of a tank etc, and consideration about how to move the tank electrodes or object to reconstruct a 3D shape.
3. Adjusting EE constraints such as adding tetrapolar differential arrangement(I'm working on this now, and also trying to source a good 32 electrode cable to spec).

Progress Update: 

EE : Updates will be easier when people can have PCB kits and I'm working on that now. There will be breakouts on the board for I2C/UART/SPI additions for things like stepper motor control and alternate electrode configurations. 

Python Dashboard: Sebastian has been doing some refactoring updates which make it easier for people to contribute, read the code and add new reconstruction algorithms. 

There is also an open request on github for better reconstruction algorithms which can be tested in the mock data mode(on the example datasets) that will be merged in shortly.

Firmware: Rob has volunteered to convert to an open toolchain, which is great and should make it easier for people to modify. 

ME and rotational tank set up: This is required to obtain a 3D model at higher spatial resolution. How should this best be done? Could it be a stepper motor controlled cylinder which could slide up and down a tank? What does this look like? Can we brainstorm ideas? Are there ME designer's on this list? 

[Eamon Egan]

Hello, group.

I am  wondering whether there is a way to determine the spatial resolution that could be achieved in a simplified ideal case, or indeed whether anyone has already done this and published the results. Perhaps with a simulation tool.

I would start with a 2D model, with the base material having a constant isotropic conductance, and a few simple objects within (for example, simulated apple seeds) having either zero or infinite (or arbitrarily high) conductance.

I would assume zero electrode-to-material interface impedance (tetrapolar measurement could deal with this in any case).

The resulting simulated impedance readings (simulated voltages when a given current is imposed) could be fed into whatever reverse-solution models are currently being used for the imaging.

This would give us an idea of the limits of resolution we could hope for with a given number of electrodes.

Anyone know if this has been done, or if not, what simulation tool would be appropriate (and, hopefully, accessible)?

Eamon

--
You received this message because you are subscribed to the Google Groups "OpenBiomedicalImaging" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openbiomedicalimaging+unsubscri...@googlegroups.com.
To post to this group, send email to openbiomedicalimaging@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openbiomedicalimaging/CAD%2BDdJwhCp9%2BkC9OZOdyzufRS05%3DCdQ8cdS3Cd%2BAjzu_hNvYgw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Jean Rintoul]

I'm aware of these toolboxes but haven't tried them.

https://github.com/liubenyuan/pyEIT

http://eidors3d.sourceforge.net/

https://github.com/astra-toolbox/astra-toolbox

The first two allow for simulations to be done and a classic simulation image seems to be the Shepp-Logan phantom. 

To unsubscribe from this group and stop receiving emails from it, send an email to openbiomedicalimaging+unsub...@googlegroups.com.

To post to this group, send email to openbiomedicalimaging@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openbiomedicalimaging/CAPkBbwtPPu8gk1_BwNutday%3DXQpzfoBf9hRp%2B6ZoDrHE6gGmJg%40mail.gmail.com.

[Daniel MJ]

Hello All,

Real late to the game here and I'm just getting started with everything EIT so I don't really have much of a leg to stand on with any opinions just yet. I'm am however a mechie so if you need a sounding board for ideas in the mean time I'm here, I have also done some work in robotic motion and specialise in inertial navigation.

Also, from what I do currently understand of EIT (which as I mentioned is hardly exhaustive) I'm inclined to agree with Eamon, I think some good insights could be made with highly controlled models, perhaps 2.5D/thick slice test cases can be made.

Daniel