Name: Democratizing Ultrasound via ML (TF Dev Summit '19)
Uploaded: 2020-03-25T08:29:14.000Z
Duration: 5 min 15 s
Description: Thousands of YouTube videos with English-Chinese subtitles! Now you can learn to understand native speakers, expand your vocabulary, and improve your pronunciation...

NATHAN SILBERMAN: Imagine you're in a traumatic accident

One of the first things an emergency room doctor may do

it to inspect your heart, your lungs, your kidney.

Ultrasound has become an indispensable window

to ultrasound is still limited by a very high price,

it's usually a large cart-based system they have to push

and then, finally, years of education and training

has no access to medical imaging of any kind,

5,000 children die every day from pediatric pneumonia,

from totally preventable complications relating

device that connects right to your smartphone.

the Butterfly iQ is a personal ultrasound device,

Butterfly's ambition is to democratize ultrasound.

The price and size of the device have solved the cost and form

But to truly make ultrasound universally accessible,

we need to solve two additional problems.

where exactly to place the ultrasound probe.

That's typically performed by a stenographer.

Now, in order to solve the access problem,

What you're about to see is our machine learning solution

to guidance called, acquisition assistance.

After indicating in the app what the user wants to image,

On the bottom, what you're going to see is the ultrasound image.

And on the top is an augmented reality interface

that shows the user turn-by-turn visual directions

that indicates how exactly to move the ultrasound

device in order to acquire a diagnostic image.

After acquiring a diagnostic image, we need to interpret it.

One interpretation model that we've developed

is for ejection fraction, an essential measurement

Ejection fraction captures the ratio of blood volume

entering and exiting the left ventricle each time

This is currently measured by clinicians by hand

and then evaluating that change over time.

Unfortunately, there's great disagreement,

even among expert clinicians, with regard

to exactly where to place those tracings or even about which

This chart shows a handful of frames and the responses

we obtained from six different expert stenographers regarding

This kind of disagreement introduces a real problem.

evaluate a model when you don't have access

we seek statistical indistinguishability.

were to show you a bunch of different estimates,

some from a machine and some from humans,

you wouldn't be able to tell the difference.

So, for example, on the left, what you're seeing

is our model in red, which is distinguishable.

Whereas on the right, if I were to remove the colors,

estimates come from the algorithm and which

We use a conventional encoder-decoder architecture.

For each frame, we predict whether the frame quality

is sufficiently clear for reliable assessment

and also produce a per-pixel segmentation

Now that each frame in the video has been segmented,

we select a single heart cycle that is of the highest quality.

Finally, we estimate the ejection fraction

using the largest and smallest areas produced by our model

Subtitles ListPlay Video

Democratizing Ultrasound via ML (TF Dev Summit '19)

bunch

access

scale

average