Closed charlesbmi closed 1 year ago
Patch coverage is 84.58%
of modified lines.
:loudspeaker: Thoughts on this report? Let us know!.
Thanks for the great new functionality, @charlesincharge! I thought the notebooks are great as they are. I'd agree with Newton on reducing some of the complexity the main DAS function would be worthwhile.
For validation/integration testing, we could add a notebook that does the comparison as you suggested (and as we discussed doing for the stimulation results as well). This could be a separate story/PR though.
Seems like the stride
installation is failing in docker-test
. Perhaps this has to do with conflicting dependencies with xarray
? Anyone have guidance on this? Perhaps we could move stride
into the poetry
dependencies?
Thanks @charlesincharge! The notebooks look good to me, thank you! I made some minor changes to keep it consistent across the two, but feel free to revert or change something I missed or is not correct. Could you transfer them to a python file so they're rendered in the documentation gallery, please?
Regarding adding stride
into poetry
-- I believe that causes an issue with the pip
deployment of the package; @NewtonSander, can you confirm?
Thanks @d-lucena ! Also, looks like the tests are now passing; I wonder if the stride
installation issue was related to https://github.com/agencyenterprise/neurotechdevkit/issues/149 and now fixed.
Github Issue: https://github.com/agencyenterprise/neurotechdevkit/issues/109
Introduction
In addition to neurostimulation, ultrasound is frequently used for imaging as well. The most common type of ultrasound imaging is diagnostic/anatomical ultrasound imaging, sometimes called B-mode imaging. We add B-mode imaging functionality and examples here: to provide teaching materials, showcase the flexibility of NDK, and inspire potential future directions.
Changes
Ultrasound imaging consists of 3 steps:
Parts 1 and 2 are already implemented in NDK; we add
demodulate.py
andbeamform.py
to implement the image reconstruction steps.We also update the
ndk.scenario
module to enable receiver locations (not just at the wavefield grid points).Behavior
New functionality. Tests should probably be added for the new libraries. We can compare against MUST as a reference implementation, similar to: https://app.reviewnb.com/agencyenterprise/bci-experiments/blob/master/experiments/neurotechdevkit-109-ultrasound-imaging/109-beamforming.ipynb/. Open to any other ideas!
Note: currently checking this in as a Jupyter notebook, because that seems easier to attach output to and view. Will repoint this to a
docs/examples
Python script after receiving feedback.