Player is loading...

Embed

Copy embed code

Transcriptions

Note: this content has been automatically generated.
00:00:00
just how much work is it to da like the two totally new domain i would like to
00:00:04
use it for medical imaging what other types of projects that it would be best adapted for
00:00:11
so the easiest i mean in my opinion the that we we have amateur
00:00:14
people so that's the best the easiest way but there's some general
00:00:19
so for be performed actually easier to use than baseball i would say in the sense because ease of
00:00:24
use graphical envision you can if your if you use signal processing and machine learning in your experiment
00:00:31
that basically it's the same work as a as we are doing because you just that you know
00:00:35
you put the signal processing blocks like uh you would do something with your images all or something
00:00:41
then the output of the the features of vectors or something and then you
00:00:45
learn or you break some kind of model and we provide certain
00:00:49
and the blue underlined look like g. m. m.s and things like that
00:00:52
we're we're going to transfer a neural network support as well soon
00:00:56
so you you build your you pray to model you know block say i'm going to train
00:01:01
and then i this block gets the features in the factors and then outputs a model
00:01:07
and that you it's you know you save it as a format i mean it's
00:01:09
all same as what we've ever so in that sense actually quite generic
00:01:14
the only what is it limiting is the what type of wonder line tools you are provided with your
00:01:18
crypt was like what i mean uh uh can you extract these type of features that you want
00:01:24
if not then you need to write that like python type of court
00:01:28
or nothing now something else is supported what adequate support it see
00:01:36
oh
00:01:38
well the human binaural that's very good
00:01:44
a
00:01:49
yeah
00:01:54
yeah okay that's very easy because sometimes in in some domains there's like
00:01:59
really the stack of tools that yes there is another domains it
00:02:03
it's very very high anything that could be like a show about a doctor containers are supported i think that's
00:02:09
something that it's not yeah but coming i think it's very soon a yeah it's almost there so
00:02:17
so yeah it's quite flexible in that sense um you have free processing power that is also important in any other questions
00:02:24
on but otherwise i think it's i think they were pretty much in time so we have half an hour copy break now
00:02:29
copy should be outside so uh use the time to discuss and then we'll be back here it and i think you

Share this talk: 


Conference Program

Welcome
Sébastien Marcel, Senior Researcher, IDIAP, Director of the Swiss Center for Biometrics Research and Testing
March 24, 2017 · 9:17 a.m.
473 views
Keynote - Reproducibility and Open Science @ EPFL
Pierre Vandergheynst, EPFL VP for Education
March 24, 2017 · 9:20 a.m.
250 views
Q&A: Keynote - Reproducibility and Open Science @ EPFL
Pierre Vandergheynst, EPFL VP for Education
March 24, 2017 · 9:54 a.m.
Reproducible Research with Bob and the BEAT Platform
André Anjos
March 24, 2017 · 10:03 a.m.
230 views
VISCERAL and Evaluation-as-a-Service
Henning Müller, prof. HES-SO Valais-Wallis (unité e-health)
March 24, 2017 · 11:35 a.m.
109 views
Q&A - VISCERAL and Evaluation-as-a-Service
Henning Müller, prof. HES-SO Valais-Wallis (unité e-health)
March 24, 2017 · 12:07 p.m.
Making experiments on remote heart-rate measurement reproducible
Guillaume Heusch
March 24, 2017 · 12:30 p.m.
149 views
138 views

Recommended talks

Team Academy
Antoine Perruchoud, coach de la team academy, HES-SO
March 15, 2022 · 1:35 p.m.
158 views