►
From YouTube: Unimatrix members conference call 2020-06-03
A
A
B
A
Do
you
mean
updates
from
yeah?
No,
we
are
in
the
process
of
handing
over
everything
to
linux
foundation
and
we
have
a
website
also
that
we
are
going
to
launch
very
soon
so
we're
just
reviewing.
We
have
a
proposal
and
a
design
and
staging
site
setup.
So
we'll
just
have
a
look
at
that
and
yeah.
It
will
be
online
soon.
B
B
A
B
B
A
B
D
Yeah,
hello
hi:
this
is
chanthang
from
umbrella
shanghai.
Hello.
Can
you
hear
us
okay,
so
this
is
general
from
umbrella
shanghai.
So,
right
now
we
are
working
with
axis
on
some
projects
about
the
uni
matrix
support,
so
we
don't
have
much
knowledge
about
the
uni
matrix.
I
think
we
need
to
learn
more
things
from
this
meeting
or
learn
some
background.
B
So
also,
we
would
like
to
open
source
the
so
we
have
open
sourced
our
inference
engine,
which
is
the
lara
engine.
So
we
would
like
to
add
umbrella
back
end
to
to
that
one.
So,
but
in
order
to
be
able
to
to
to
put
it
open
source,
there
are
some
libraries
from
ambarella
that
has
nda
right
now,
so
it's
not
possible
to
open
source
the
whole
stack.
B
D
B
F
B
G
B
E
E
G
E
E
E
E
E
E
And
I
noticed
and
when,
when
I
when
I
tried
using
it,
I
noticed
the
drawing,
it
is
difficult
to
express
relationships
between
tasks
and
it
is
difficult
to
understand
what
the
task
is
for.
So
if
there
is
a
more
excellent
method
or
useful
tool,
please
introduce
it
and
that's
all
for
task
management
experience.
I
E
G
J
J
J
B
B
J
J
You
can
do
a
lot
of
very
nice
things,
including
email
subscriptions
and
so
forth.
The
one
advantage
that
we've
found
to
using
the
kanban
kanban
on
github
is
that
you
can
quickly
turn
your
your
cards
into
issues
and
so
there's
basically
a
quick
workflow
that
says
you
know
a
project.
For
example,
a
project
manager
might
create
a
card
for
a
specific
task
and
then
the
developer
team
could
turn
that
card
into
one
or
more
issues
on
individual
repositories,
with
very,
very
low
effort.
J
G
B
We
we
need
some
help,
but
we
can
be
company
one
of
course,
but
for,
for
example,
opencv
subset.
We
would
like
to
discuss
with
the
other
companies
what
what
their
requirements
are.
So,
if
there's
someone
willing
to,
I
mean
for
creating
the
docker
base
layer
with
opencv,
we
need
to
cannot
configure
opencv
the
right
way
for
the
subset
that
we
want
to
have
there.
G
E
E
E
G
G
E
L
G
B
C
B
B
C
K
G
G
E
G
B
G
G
B
B
G
G
L
E
B
J
J
J
J
J
Many
of
you
have
already
seen
this
slide
in
our
conversations,
but
you
know
who's
participating
in
the
teeny
working
group,
really
it
kind
of
started
with
myself
saying
hey.
I
think
this
is
important
and
taking
some
initiative
and
access
providing
a
nod
and
saying:
yes,
we
think
it's
important
too
and
now
the
teeny
working
group
is
really
looking
for
for
members.
So
we
would
really
like
to
have
someone
from
each
of
the
you
know:
the
court
sponsor
teams
participating
in
the
working
group
and
providing
guidance
and
oversight.
J
It's
you
know
the
core
unimatrix
team
on
the
one
side,
who's
developing
pieces
for
the
unimatrix
stack
itself,
the
unimatrix
consumers,
ultimately
helping
them
to
test
their
applications
on
the
unimatrix
stack
and
finally,
for
the
camera
oems
to
help
them
be
compliant
with
the
stack
and
so
really
it's
a
question
of
which
of
those
a
b
or
c.
Do
we
prioritize
right
now
in
2020
and
ultimately
I
think
that
it
should
be
a
the
unimatrix
contributor.
J
Finally,
in
the
short
term,
we
would
like
to
deliver
a
clear
proposal
for
basically
a
charter
for
the
working
group.
In
the
medium
term,
we
want
to
basically
deliver
a
recommendation
for
one
or
more
of
the
toolkits
for
doing
tests
and
evaluation,
and
then,
in
the
long
term,
we
think
that
there's
a
real
opportunity
here
to
you
know
to
actually
implement
some
new
technology,
some
new
toolkits
that
address
the
key
problems
that
the
unimatrix
stack
attempts
to
you
know
to
to
to
eliminate
in
the
technology
space.
J
Clearly,
you
know,
unimatrix
is
an
important
modern
project
for
a
really
critical
business
vertical
and
with
you
know,
tne
I
think
has
to
be-
has
to
be
modernized
around
it
as
well.
So
I
think
you
know
actually
implementing
some
new
modern
technology
in
these
toolkits.
This
would
provide
a
really
great
test
case.
Unimatrix
would
provide
a
great
test
case
for
for
getting
that
done
so
far
we
in
may.
J
We
had
really
detailed
discussions
with
panasonic,
I
pro
hikvision
and
tahuah,
and
in
june
we
want
to
continue
those
discussions
with
the
other
unimatrix
teams,
and
the
primary
goal
is
to
at
least
in
june
is
to
find
a
way
to.
You
know
be
useful
for
the
teeny
working
group
to
be
useful
during
the
hackathon
and
then
the
secondary
goal
is
for
us
to
submit
this.
J
I
think
maybe
another
thing
to
think
about
in
with
regard
to
the
hackathon
is,
you
know:
are
there
pieces
of
the
unimatrix
technology
stack
that
are
that
are
ripe
for
tne
already
and
if
they
are,
then
we
want
maybe
to
refocus
on.
You
know
getting
an
implementation
of
tne
for
those
pieces
of
the
stack
sooner
rather
than
later.
So
as
a
general
rule,
as
soon
as
something
becomes
testable,
we
would
like
to
start
testing
it,
and
that
is
really
kind
of
the
way
that
the
teeny
working
group
will
show
its
show
its
value.
F
Thank
you,
brian
just
one
thing
to
mention
this
is
that
confidence
we
are
going
to
be
following
up
as
well
directly
with
the
linux
foundation,
just
around
sort
of
how
to
best
structure.
That
proposal
for
for
that
and
how
focal
as
an
organization
specifically
can
you
know,
can
support
that
in
the
the
best
way
to
the
best
way
to
do
that,
so
we'll
have
some
feedback
from
them
as
well.
J
B
J
B
It's
it's
not
complete
if
you
compare
to
on
with
it's
more
like
a
draw
very
draw
thing,
so
I
I
just
just
wanted
to
get
some
feedback,
so
we
don't
have
to
discuss
it
here,
but
it
in
the
private
repository
there
is
a
a
file
called
scene.proto,
so
it's
a
protobuf3,
basically
a
an
interface
for
or
a
protocol
for,
transferring
a
state
of
objects
like
it
on.
If
you
have
a
a.
K
B
B
B
So
let's
start
with
the
presentation
so
from
axis
side,
we're
going
to
do
just
an
overview
presentation
about
the
goals
and
so
on.
Also,
maybe
panasonic
can
fill
in
with
the
grand
design
after
that
and
then
we're
going
to
have
an
update
about
the
inference
engine
and
we
probably
gonna
have
an
update
about
opencv.
B
Also.
We
will
maybe
present
something
about
this
protobuf
api
that
I
described
now
and
then
there
is
a
presentation
about
budget
like
a
resource
budget.
So
last
time
we
we,
I
think
they
were
both
from
panasonic.
I
pro
high
vision
and
axis.
We
want
to
kind
of
all
give
us
our
currents,
the
lay
of
the
land.
So
we're
also
going
to
do
a
presentation
about
that.
Yeah.
B
B
B
B
B
So
my
suggestion
is
to
use
the
time
from
12
utc
to
8
utc,
at
least
on
the
first
day.
I
think
the
second
day
it's
up
to
yourself
how
you
want
to
do
when
you
want
to
do
the
hacking
and
so
on,
and
then
on
the
third
day.
I
think
it
would
be
a
half
day
for
just
presenting
the
results
of
your
hacking
tasks.