Manifund foxManifund
Home
Login
About
People
Orgs
Categories
Newsletter
HomeAboutPeopleCategoriesLoginCreate
🦑
🦑
Venkata Subhash Chandra Sadhu

@subhash-sadhu

51

$0total balance
$0charity balance
$0cash balance

$0 in pending offers

Projects

AI-assisted low-cost ultrasound scanner
By Venkata2Following1
AI-assisted low-cost ultrasound scanner
🦑

Venkata Subhash Chandra Sadhu

23 days ago
Progress update

What we did

  • Built a benchtop experiment as proposed, using off-the-shelf components: a piezo ultrasound transmitter, a tissue-mimicking phantom.

  • Generated ultrasound waves using a non destructive testing transducer repurposed for our use case; this was the easiest to acquire and the simplest way to get started. We had to be careful to not exceed the component specifications and unintentionally destroy the transducer, and had to tweak our experiment design to accommodate that.

  • Used high-voltage electronics to drive the transducer. This was a steep learning experience; we learned that the typical off-the-shelf driving circuits are not designed to accommodate our creative usage and self-destructed in interesting ways and ended up in some (literally) hair-raising debugging experiences.

  • Went back to the drawing board and rethought the driving electronics, and managed to figure out a very different solution by repurposing high-frequency power electronic circuits.

  • Tried imaging an old second-hand imaging target (budget acquisition) and, we realized that that was an unnecessarily complicated starting point for validating our design. We then switched to simple lab-made phantoms with metal wires serving as acoustic reflective structures immersed in water.

Upshot

We have progressed from a theoretical concept to a working, integrated hardware system and successfully completed our first phase of experimental validation.

  • We designed and constructed the hardware necessary to empirically demonstrate the sensing approach.

  • One of our biggest technical uncertainties was whether environmental acoustic and mechanical noise would overwhelm the extremely small signals we are trying to detect. Our experiments demonstrated that our sensing architecture is robust to such noise even without specialized mechanical isolation or active stabilization – which is very promising for building a wearable device in the future.

  • We have developed a digital twin system model to capture the behavior of the sensing architecture. This is helping us make stronger design improvements.

Next steps

Our next goal is to demonstrate end-to-end ultrasound imaging by extending the sensing architecture we have validated. This will involve refining our hardware,improving our digital twin, and testing the imaging algorithms that we have implemented. Our demonstration will focus on generating ultrasound images of fruits immersed in water. These biological samples provide a useful test environment for evaluating imaging performance with tissue-like targets.

AI-assisted low-cost ultrasound scanner
🦑

Venkata Subhash Chandra Sadhu

about 1 year ago

@acx-grants Thank you so much. This will go a long way in helping us bring this to the market.

Transactions

ForDateTypeAmount
Manifund Bankabout 1 year agowithdraw23000
AI-assisted low-cost ultrasound scannerabout 1 year agoproject donation+23000