►
From YouTube: Applying IPLD to Medical Imaging - Chris Hafey
Description
Check out this amazing real world use case for IPLD + IPFS! Chris Hafey explains how content addressability is shaking up a world of muteable data: the healthcare industry.
For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS
Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5
A
My
name
is
Chris
heavy
and
I've
been
working
on.
I
PLD
to
the
medical
ng
domain,
I've
been
working
in
medical,
imaging
and
healthcare
for
most
of
my
career
over
20
years
now,
and
really
excited
about
moving
from
a
field
which
is
primarily
about
mutable
data
into
the
world
of
content,
addressed
data
and
obviously
that
brought
me
to
the
idea
thus
protocol
labs,
I,
peeled
ebrill.
So
just
a
couple
of
quick,
slides
I
have
through
to
the
other
year.
A
So
one
in
terms
of
background
or
so
in
medical,
imaging
and
in
healthcare
standards
are
really
important
and
the
standard
for
medical
imaging
is
called
DICOM
and
basically,
what
it
does
is
it
takes
a
given
image
like
a
CT
scan
or
MRI
or
an
x-ray,
and
it
puts
it
in
one
file
and
in
the
file
at
the
very
beginning
of
file.
Are
things
is
a
header
which
has
like
patient
name
date
of
birth,
gender
and
then
the
pixel
data
and
I?
A
Think
there's
been
you
probably
heard
of
people
using
my
PFS
for
medical
imaging
in
the
past,
but
there's
a
real
big
problem
with
it,
and
that
is
this.
Header
does
tend
to
change
over
time
kind
of
breaking
the
whole
unity
content,
Dressel
model
that
we
would
want
with
I
can't
fast.
Now
we
actually
want
healthcare
too,
and
so
a
couple
examples
when
that
can
happen
is
if
you
get
married
or
you
change
your
name
for
some
reason
that
can
trigger
basically
a
rewrite
of
these
files.
A
A
There's
a
lot
more
technical
depth
of
details,
I've
actually
prototypes,
that's
using
GSI,
PFS
and
it
works
is
pretty
cool,
but
I
want
to
kind
of
communicate
the
high
level
things
about
what
I've
discovered
or
what
I
do
and
also
collect
feedback
people
have
it,
but
basically
splitting
the
DICOM
file
into
two
and
storing
it
directly
in
IP
LD.
It
turns
out
there's
a
if
you
haven't
gotten
into
I
killed
me
too
much.
A
If
you
take
this
and
you
go
like
into
healthcare
in
general,
here's
like
a
typical
healthcare
data
model.
It
is
inherently
graph
based.
You
have
a
patient.
You
know,
there's
demographics
up
here,
maybe
I'm
multiple
image
studies
they
have
taken
over
carry
of
their
life.
Various
encounters
observations,
reports,
lab
results
and
machine
learning
results,
and
so
this
is
actually
kind
of
taken
from
the
healthcare
or
standards
space
right
now,
in
terms
of
how
we
think
about
things,
and
what
I'm
looking
at
doing
is
how
can
I
apply
that
in
IP
LD
world?
A
So
essentially,
what
it
kind
of
requires
is
assigning
kind
of
like
grouping
these
different
pieces
of
data
together
into
so
they
can
be
kind
as
a
single
block
or
single
CID,
and
so
there's
some
natural
kind
of
break
points
to
connect
the
existing
healthcare
data
model
into
the
IPL.
Do
I
build
the
world
by
assigning
C
IDs
different
points,
and
you
can
see
I
also
added
the
concept
of
a
root
appear.
A
This
gives
me
essentially
change
tracking
and
a
longitudinal
sense,
whereas
changes
the
patient
record
occur,
I
can
record
that
and
then
basically
generate
a
new
root.
So,
in
this
case,
I've-
maybe
change
something
in
the
patient.
Demographics
that
the
name
is
changed.
I
generate
a
new
block,
I
peel
the
object,
if
you
will,
with
a
new
C
I,
need
a
new
group.
I
can
reuse.
Basically,
all
the
existing
data
that's
already
stored
in
IPL
D
and
in
is
the
pixel
data
which
is
way
down
here.
The
bottom
image
printed,
so
images
are
huge.
A
They
count
for
something
like
80%
of
the
storage
in
a
hospital
and
there's
huge
benefits,
making
immutable
and
not
even
deal
with
changing
the
data,
but
basically
making
it
reusable,
and
not
only
within
just
general
clinical
care
for
healthcare,
but
also
some
new
things
which
are
emerging
like
machine
learning.
An
AI
and
image
sharing
using
on
the
top
is
open.