Press "Enter" to skip to content

The robot doctor will see you now: evolution of robotic vision | Anjali Jaiprakash | TEDxSydney


the power the beauty and the elegance of
vision a true gift of nature life and
evolution I’m a biologist working with
robots and as a biologist I love the
study of life and the impact of this
ever-changing environment on our
evolution and by embracing and adapting
the positive features of technology our
world and future is only limited by our
imagination welcome to my world of Robo
biology my focus area is healthcare
which I believe is a basic right my
vision is to create a world in which
robotic technology enables affordable
medicine for all to get there I’m going
to explain three things where are we at
with robots how I use their robotic
efficient technology in healthcare and
how it’s all going to work together to
start let’s think about robots and let’s
go back in time about 542 million years
ago there was a burst in evolution where
we saw the biggest diversity of life
happened at a relatively short period
called the Cambrian explosion so simple
organisms evolved into more complex
beings because of change in environment
no one really understands the reasons
behind this change but it was believed
to be because of change an oxygen change
in temperature and most importantly the
evolution of fission so organisms were
able to see for the first time and this
led to a whole new creative way of life
from looking for food
to predatory behaviors
and the different sexual patterns we’re
seeing the same thing in robotics now
there’s a perfect alignment in changing
technology with great computing power
big data and ER evolution an AI is
different to robots yeah so AI is the
brain better than ours at pattern
recognition and connecting the dots
that’s required for precision decision
making but having said that AI is also
not as sophisticated as our brains in
terms of perception but we are getting
there and this physical thing that
understands this decision to act is the
robot we humans do this beautifully
because we sense the world we touch we
hear we feel and see and even if we
cannot see we learn to live and adapt
and move forward robots are not there
yet they’re dumb God can’t call them
dumb let’s call them they’re like little
children influenced by the evolving
technology and what we teach them and
they’re just learning they’re learning
had a sense in pennyways and we believe
that vision in robotics could be one of
the most efficient and effective ways of
unlocking their potential to interact in
our world we at the center of robotic
vision create robots that see better so
it can sense better and eventually
interact with complex unstructured real
world environments so what is the
difference between the way humans and
robots see there are nine ways humans
perceive the world through vision and
how the brain and the vision works is
really complex but it’s beautifully
coordinated that allows us to see
and interact the way we do we cannot
train our robots to see and understand
like the way we do but we have moved way
ahead from robots seeing nothing –
robots with a single camera and my
complex ways of sensing our environments
one of the ways that robots can see is
using a computational imaging technology
called as light fields light field is
all limb innocence it is the collection
of light rays in every direction and at
every point in space light field imaging
is recording the light field so how does
it work each of you here are looking at
me like individual cameras together and
collectively you’re acting like a light
field camera each of you are also
looking at me and a slightly different
perspective so your focus on me is from
that perspective so some of you here can
see me very clearly and some kind maybe
and I because of a tall person in front
of you or big hair but together and
collectively you can see every inch of
me and if I’m recording all these views
and depending on where I want to focus I
will choose my views I will choose views
from you to get the information of my
palm and even those lines I will choose
views from you to get the information of
the back of my hand
or I will choose all these views to get
a 3d view of my hand roboticists around
the world are using this technology to
understand our surroundings a lot more
this allows robots to work out distances
how to look beyond obstacles and even
understand textures this ability
is allowing robots to evolve and see
better for example in the military
they’re using this technology to
navigate and search for missiles in the
virtual and augmented world we are
creating enhanced experience for movies
and gaming light-filled is radically
changing the lens through which we see
the world I’m interested in healthcare
so we’ve used like field technology to
build a device that can image the retina
and it’s going to look something like
this we’ve named it the retinal plan
opto scope but before I go on and
explain how it works let’s understand
some of the challenges behind retinal
imaging the retina is the back of the
eye the eye is also a living camera so
it’s really hard to take an image of an
imaging system it’s like I’m sure some
of you here are guilty of this but
really it is like trying to stand in
front of a mirror and take a picture of
yourself with a flashlight on like that
one there I love that picture to see the
retina you also have to see through the
pupil it’s like trying to see the world
through a keyhole
and that’s why specialists sometimes use
this drop to dilate the pupil to make it
bigger so the entire process of taking
an image of your retina could be about
so why the retina it allows us to see it
is our evolutionary advantage and it is
the most efficient and effective way of
interacting with the world around us so
percent of eye disease leading to
blindness can be prevented if it is
monitored regularly and treated in time
there are 285 million people suffering
from preventable eye disease modern
retinal equipments can provide good
quality diagnostic images but these
numbers say something is wrong there are
some really great efforts out there
trying to reduce this number but one of
the main challenges that remain is still
how to take a good quality diagnostic
image last month FDA approved a set of
AI algorithms that can detect a retinal
disorder this is great news this means
that automated precision diagnosis is
already possible but how am I going to
take this image current precision
equipment look something like this they
are large they are bulky they are
expensive and even if you can afford it
you try and drive it back on the outback
on these rocky roads you may need a
technical specialist or an expert to
recalibrate it or fix it you also need a
trained specialist to take an image of
your retina this device here attempts to
solve these problems we capture 2600
views using a single shot the optics of
this device is really complex but the
device itself is quite simple really
so it makes it realistically portable
because we focus after taking an image
anyone can take a precise image we also
offer good quality diagnostic images
because remember in light field your
not missing any information and you’re
choosing your views depending on where
you want to focus the idea really was to
offer new diagnostic features way beyond
any current retinal equipment that you
can find for the first time 3d depth
maps and metric measurements of your
retinal features that is patient
specific so you can track and monitor
changes that is specific to you and all
this without dilating the eye so what
does this really mean a device such as
this combined with an intelligent brain
like the one the FDA is approved could
truly mean non-invasive preventative
personalized eye care anywhere research
also shows that our retinas holds the
potential to predict cardiovascular risk
and understand the health of our brains
by understanding diseases like
Alzheimer’s and Parkinson’s the impact
of this will be far-reaching for our
health and maybe one day we will be able
to use all this information with the
robot to act to act and perform
autonomous intervention
this means robots will evolve they will
see the world and they will think and
they will be able to act this leads me
to my final concern how is this all
going to work together we’re in an
exciting time of evolution in the world
of technology that could lead to
exponential improvement in capabilities
but this is not enough there’s a lot of
science and technology out there but
bringing this to the people who really
need it the most is the biggest
challenge in healthcare and we can do
this we can do this by understanding
people their motivations their behaviors
and their needs we must always put
humans first to gain insights on how to
design our technologies and services so
it reaches people and let’s not forget
inform policies to ensure society’s
needs we live in an age of disruption
things are a lot easier but at the same
time more and more complex technology
has disrupted a number of fields from
hospitality financial services shopping
healthcare is that last great frontier
to innovate the way medicine is
practiced and delivered this will not
just have an impact on developed health
robots are evolving similar to the way
we are we did we are the most successful
species and we are the creators of this
robotic world so it is our duty to
ensure that they serve the greater good
for me this means transforming
healthcare so every individual can
augment their lives and benefit from
this new form of robotic life and I see
I see a future where humans and robots
are all working together to benefit us
this is my world of rubber biology thank
you
[Applause]
[Music]
[Applause] [Music]
Please follow and like us: