0:00:13it nice but the
0:00:14yeah speech can kind of thing but choose by see image and signal processing
0:00:19and we have here and
0:00:20for as the have this can who twenty
0:00:25and that are the hot topics
0:00:29and two oh so as to the community at large of signal processing
0:00:35so um
0:00:37a for the yeah i
0:00:40and that the
0:00:42uh about a
0:00:46is to be a a
0:00:48first that which is a the perspective of the scene or the best G
0:00:52to understand the can the kind for five
0:01:03a i like the
0:01:07and the prediction some genes and proteins and a is to be able to it
0:01:11is is the conditions
0:01:14and the prior to choose
0:01:16but G
0:01:21the band
0:01:22and the size of a
0:01:23down to the size of a
0:01:26the the content was a scale and by different that from for example a pad this is medical imaging
0:01:34the the the some of P be to go just to be
0:01:41she's which i produce a information which just produced
0:01:44we have a number of topics which i uh i have uh come to the two
0:01:50and discipline
0:01:51which i a
0:01:53to to this process and and then said
0:01:56good friend mathematical gym which is how can put
0:01:59some i think so that's processing
0:02:10i information a set uh but and then
0:02:15to to of us steps
0:02:19processing knows those problems
0:02:23can can and then
0:02:26to a a a a time of just first we choose to put into a just uh a just a
0:02:32images which type of crime
0:02:35so that the the fact to extract information from the background the relevant information
0:02:41to the classification of information
0:02:43and the to be able to put in in um
0:02:51in this manner under images are kind of
0:02:55in this manner as
0:02:59and and uh i and in and then
0:03:03no side then i
0:03:07i just to uh for size
0:03:11and the fact that for methods
0:03:15for the processing session
0:03:18uh for the
0:03:21to look
0:03:23a can to the to the information
0:03:27to choose so uh uh uh a type is for example
0:03:33compressed sensing is becoming
0:03:35a of research
0:03:37but had to go
0:03:39because the the
0:03:46to to spend to much time
0:03:51and a a which is how does from
0:03:56it it's also about uh
0:04:01type of research
0:04:07so so i could use a set have two
0:04:11so that
0:04:12was which i X uh
0:04:22to test uh
0:04:28a uh a a uh and each of them
0:04:32so so as to have this
0:04:40a process right
0:04:43which the channel
0:04:44have to the go to a a a uh of the um a
0:04:51as the beaches
0:04:56yeah uh i use and try and
0:05:03can give some idea is also at could me find find
0:05:08G Q
0:05:25i'm can some stuff
0:05:27so that's think that's that uh
0:05:31so are so T R screen D that's as far as E and a go to note
0:05:40access from shown
0:05:42and information and uh
0:05:51my first slide okay
0:05:56and T
0:05:58because i asked to at some found notion and that function as a cat and so i don't you can
0:06:09i have
0:06:11i have a sounds
0:06:13have stop to learn and then sit products and i've that's that's you know
0:06:21and and so
0:06:27so that so that
0:06:37and then suppressed
0:06:40so um
0:06:42and that X M M M and that's consistent
0:06:51that's as a
0:06:52so that's that's so
0:07:00and so
0:07:02and at the i-th sensor and this is yeah
0:07:09that sounds
0:07:12a uh
0:07:16thanks to that construction
0:07:19five so
0:07:21so i don't function and sabrina
0:07:25and uh i know that have to reconstruct
0:07:30i have
0:07:37the presence of a
0:07:39so here
0:07:42and that's that's faq
0:07:49so that's accent
0:07:52i have to say that the construction masters channel
0:08:01i don't
0:08:02i have a pension plan
0:08:06oh doesn't that
0:08:10so um
0:08:12the acquisition system
0:08:16and and twenty seven are expressed as a function
0:08:23and at times i was
0:08:27of uh uh
0:08:29optimization is convex and how
0:08:33a a a a a a and uh some kind of
0:08:38and uh
0:08:40or section
0:08:46a and have and it's fast but was that because we've got a a a a a a three D
0:08:51the mountains to kind
0:08:56press sound system so it's down
0:09:02so a five a script
0:09:06and oh oh can script
0:09:10yeah that's a
0:09:19uh i i stands as a facts phones
0:09:24i have some
0:09:27and and that's a that's it's function
0:09:33that that
0:09:37facts phone so so that yeah
0:09:41the and that and a and the convolution "'cause" of times perhaps function as as nine
0:09:47that's a concepts
0:09:55a she used on the point spread function
0:09:58so that an absent system a
0:10:03so account the mouth
0:10:06i have just stash um
0:10:11a a i have a complex to that so that the structure i
0:10:19and uh that's a constant
0:10:24uh i
0:10:29i was just a concept for example of god
0:10:34construct estimate costs
0:10:36for five example some kind of cats watson's mention that
0:10:44as a last step to
0:10:48for some sequences come out
0:10:52match well
0:11:07and so on
0:11:09have come to that
0:11:12to the concept of a complex when the structure
0:11:16that's that's a problem
0:11:18oh have
0:11:21uh statistical stuff such a complex task
0:11:26and five is not uh
0:11:28to define approach to insist substance and not rocks
0:11:33and uh have questions that is the kind of news and most of them to a convent but sectors
0:11:41in some sense
0:11:48we propose is that we
0:11:50uh_huh go through an addition that the have ten minutes for discussion open discussion on all the the the topic
0:11:58so yeah also go on was biological imaging
0:12:02and perhaps web addresses a little for being very focused some biological imaging because this was also andrew
0:12:09the supposed to to about medical imaging that in
0:12:13uh uh in through a right now
0:12:16so yeah
0:12:18hmmm and it is so was so as so he of the challenge signal processing challenges for
0:12:26and uh uh it has to do with the fact that this three D uh plus the and the fast
0:12:30so that was
0:12:31really thought telling you about the convolution know as it i mean as a a a uh than men's the
0:12:38i mean if they can be
0:12:40that's a can be then because um P T
0:12:43satisfactory this factor
0:12:45to to have a a a a a a lot of the challenge is and you have the set of
0:12:50image analysis because
0:12:52this is a a and the by just to have numbers
0:12:55and the so the a set so the to uh uh my west image shows which have to do with
0:13:02because because tracking so this self worse more those in images
0:13:06looking that's that was also a
0:13:08i have the show if the gene expression
0:13:11for a has and and you know some the some of the problem of that that to the
0:13:15a a a a a and to for for image processing
0:13:18so i
0:13:20i i i i will actually this is the sort of give you some example state of the art
0:13:24and taking a a uh algorithms that have been developed a people in a common T that she i been
0:13:31used by religious
0:13:32and i mean this is this that of the to the the challenges that i to go on the and
0:13:37i with
0:13:38sure but
0:13:39for example here
0:13:41okay so if if
0:13:42you see here you have
0:13:44this is an example of a uh uh you know time laughs my first P
0:13:49where uh uh actually this is a a ms so so
0:13:52uh uh uh the end of the proposed um that has been
0:13:55the the the uh is inside the nucleus and have
0:13:58the are just like to
0:14:00to is month so for of those from more
0:14:03and so that was a better here is the tracking i
0:14:07uh that will actually five
0:14:10in so i'm and and i is to do that
0:14:13if are less fashion using
0:14:15oh to defined as function using the fast
0:14:20i mean this is
0:14:21uh just works very well for
0:14:23a single party
0:14:24but the
0:14:25i'll the challenge
0:14:27is uh
0:14:28you may have this type of
0:14:31so you have
0:14:32yeah yeah of those party
0:14:35oh have of for and the background
0:14:37and so uh a real challenge
0:14:41for for for for that
0:14:42data so it's
0:14:43you know
0:14:44you'll three
0:14:47a uh so next challenge
0:14:50uh so it has to do with
0:14:52uh and uh uh so uh shape and
0:14:55you need it
0:14:55so he it is simple
0:14:59so the violent adjust here this is the uh
0:15:02face got five
0:15:04my cross is so
0:15:05yeah i i is to outline of cells
0:15:08not just for counting them but because we want to extract the gene expression profile
0:15:13it's so this is also like can of it
0:15:15the ball of that's
0:15:17by all
0:15:18kind of like
0:15:20a fact let's
0:15:22and then if you uh uh
0:15:25like that of
0:15:30so that what what you have
0:15:31principle here
0:15:35so what's important here it's
0:15:37for all those cells
0:15:39with time so why would you want to do that
0:15:41because all cells actually what
0:15:43right thing
0:15:45a different genes your
0:15:46that'd been through recently label and and so here you you you know the outline of this
0:15:51the the cell
0:15:52but what you would like to quantify is the expression
0:15:55oh uh i mean the amount of fluorescence in in in both cells here
0:15:59and and so get that uh time course of of fluorescence and and and and and uh this this will
0:16:05be the date so that the biologist one two
0:16:08you we uh i mean to extract so what a
0:16:11here uh the challenges i mean the challenges is very poor contrast of those images
0:16:16dealing with much more complicated it's uh shapes than those used cells which are round
0:16:21and things cell division it should using models of time evolution
0:16:25uh and doing global optimization in space and time
0:16:29dealing with crowded images touching cells in to reducing repelling forces high throughput puts mean this needs to be done
0:16:36on you amounts of cells and power that imaging
0:16:39fast row put able and the uh above all easy to use a gram so that people uh i really
0:16:45applying the mean brock
0:16:46so now uh relating more to what to a lot was telling you
0:16:50uh uh you know there this a problem of extracting features from images and and so for example this was
0:16:58a program developed
0:16:59it's the
0:17:00it's a program for tracing uh acts on of of neurons but
0:17:04at some your to me
0:17:06and and not
0:17:07i mean i mean this works well it's uh use by by people
0:17:10but uh uh uh
0:17:12i of course uh i mean life is is more complicated than that
0:17:16and what i see as a challenge an opportunity uh a of uh for for for a signal processing
0:17:22community actually coming up with
0:17:24uh what that would call key point for by you because key points have the huge six S for computer
0:17:30vision so is looking like
0:17:31points of interest in image so what we would need a by all we are not like we have three
0:17:36D data
0:17:37so we would not need for well we can first look at two D but optimize
0:17:43a for scale uh a translation rotation invariance
0:17:46also i mean what we're seeing we have background verses uh a structures of interest so doing detector that we
0:17:53might seem i
0:17:54uh suppress press background and
0:17:55probably use this kind of idea for designing wavelet type representation that can
0:18:00and has the feature of interest
0:18:02and and and and suppress the background
0:18:05but the real problems in three D
0:18:07okay because there's nothing in three
0:18:09oh O case to two minutes okay
0:18:12and and and and so i that she do that in three you
0:18:15so in three you have a interesting structure like sheets membranes
0:18:19and he the challenges developing that's a durable the vectors wavelet since really
0:18:24and uh above all be computationally efficient uh uh because i mean the are huge amounts of data
0:18:30so i just wanted to
0:18:32uh here are uh i i mean this is perhaps what where banning in the lab but just to show
0:18:36you that
0:18:37uh one can do a very interesting steerable wavelets so we here are
0:18:41i'm in in in in two D that look like has since but there a reversible
0:18:46uh where let's but the
0:18:47this seems also to be a possibility to do them in three D and this is uh
0:18:52pretty much uncharted so doing
0:18:55where it's that can what's agents
0:18:57and i yeah side like just like to have a size
0:19:00if you're working in biology so that
0:19:03it's a disciplinary research where there are lots of players of course the to just the medical people
0:19:09there's the optics the microscopy the by chemistry which is the mark and the signal processing
0:19:15and there were there was also like here
0:19:17special issue on biological imaging a few years ago but is still a good entry point for those want to
0:19:22get in uh in with few okay case
0:19:59as so uh i will
0:20:00and the law speaker go a bit more into challenges that are you merging uh especially in the medical field
0:20:07wait a little bit from the biological
0:20:10um so basically number one chance number one is still increasing data that now
0:20:16already already for treat this is going on
0:20:18uh the the simplest by medical signal is the
0:20:22electrocardiogram the
0:20:24second two dimensional signal is the the
0:20:27normal two dimensional picture
0:20:29then uh in the eighties came to three dimensional pictures
0:20:33and then the four dimensional pictures
0:20:35a three dimensional close
0:20:38uh more recently especially in the the last decades uh specially functional imaging
0:20:43on top of and apple
0:20:45imaging has become very cool
0:20:46in that case you three spatial dimensions one time dimension
0:20:50functional entity for instance uh blood oxygen nation
0:20:55three plus three dimensions basically meeting three
0:20:58a a spatial dimensions a three direction vectors uh base
0:21:02diffusion tensor imaging
0:21:05then there's uh nowadays very popular uh it's starting to really get popular
0:21:09three plus three plus one
0:21:11and dynamic uh imaging
0:21:14where you can actually a monitor or you know that the
0:21:17a three dimensional flow
0:21:19inside the ha
0:21:20and especially signal processing for
0:21:22this is still in its infancy this sense that
0:21:25able try to characterise a for text flow right
0:21:27he's uh in
0:21:28he's types of images
0:21:29but then when you have
0:21:31a i thought we kind of reached a limit at uh seven dimensions here
0:21:35but then uh this uh
0:21:37and and brain database was published
0:21:40and a brain database is basically three dimensional data dataset
0:21:43plus at time aspects where it basically for
0:21:46a my
0:21:46of mice
0:21:47where for every voxel they have twenty thousand G
0:21:50which uh
0:21:52values for gene expression show basically
0:21:54but see this as it
0:21:55you which uh uh dimensional data set
0:21:58a a which is now uh just about to be explored
0:22:01to correlate it's uh with all kinds of
0:22:04for imaging
0:22:06so this is number one the dimensionality of the data is still increasing
0:22:11and signal processing challenge
0:22:14the uh plentiful there
0:22:16seconds is that um
0:22:19uh until about ten years ago was mostly about imaging structure and it to me with ct anymore
0:22:26and function
0:22:27with for instance uh a nuclear technique
0:22:30it would highlight some aspect function
0:22:32and past decade all kinds of technique became available to also look at biochemistry
0:22:39so now we can look at diseases on a biochemical level
0:22:42and then see what happens first in biochemistry uh what goes wrong in a particular
0:22:47by chemically
0:22:48and then study effects all structure
0:22:51and function
0:22:52now the data acquisition is now
0:22:54getting so far that we can acquire these triangle structure
0:22:57function and by chemistry
0:23:00the signal processing and image processing is still
0:23:03largely to we think this on as separate entities and
0:23:06i i think there still a big challenge
0:23:08in trying to come up with in to great if
0:23:10signal processing
0:23:12it actually looks at all data at the same time
0:23:16there at very important trends uh i i think shows also that the
0:23:20few of by a middle image and that was is is you ring
0:23:23at at this point in time
0:23:25used to be the case that you could get away with a validation of a segmentation algorithm for instance
0:23:30just by citing some uh uh uh accuracies that others have achieved
0:23:35and then uh a benchmark your algorithm
0:23:37to that but the problem with that is that
0:23:39all those those uh
0:23:41results published in these papers
0:23:43they are usually done different patient different test state that different test protocols different error metrics different gold standard everything
0:23:50is if
0:23:51what you see and this make it very difficult to really make an object is mark
0:23:56and especially in the pattern recognition community there's already a lot of standardised databases we
0:24:01known classification result
0:24:03we see now
0:24:04in the medical image processing field
0:24:06more and more
0:24:07words we objective quantification they make fable
0:24:11a centralized data
0:24:13evaluation scripts are run sense really
0:24:15and everybody can benchmark their data their algorithm against that particular that's it
0:24:21i i think this trendy so input
0:24:23and that
0:24:23nowadays if you try to publish uh and algorithm on
0:24:28a topic which has already been that smart
0:24:30in one of these
0:24:32a you cannot get away without
0:24:34uh benchmarking marking it against it
0:24:36and there's kind of a default rejection
0:24:38if you do not include validation experiments on these standardised uh um
0:24:43data bases
0:24:44so very important field of of of yeah it shows that the
0:24:48the the the field is but your in a that's
0:24:52for if very important new topic uh is a key to more data analysis
0:24:57right the main question to answer is
0:24:59something is changing but what
0:25:02now and you can imagine that
0:25:04if you look at an in fit individual patient uh yeah it still manageable
0:25:08uh you look at follow up
0:25:09uh is
0:25:10easy is getting worse yes or no
0:25:12uh but nowadays one tends to look at
0:25:15groups of patients anyway between a two
0:25:18and one thousand
0:25:20this study D V disease each development in two weeks writes treatment affect
0:25:25even more
0:25:26data is acquired
0:25:27normal subjects there and now you huge
0:25:30population studies right there
0:25:32every year scanning people healthy people
0:25:35and they're is getting every here and then basically wait until people get sick
0:25:39at some point in time
0:25:40and then they can backtrack
0:25:42through the date that whether they can find
0:25:44early signs of that
0:25:45you can imagine that that finding these are signs is really a needle in a a a a stick
0:25:50and uh well there's a huge uh image processing challenges
0:25:54trying to really combine all that data and to backtrack to bit data
0:26:00well another trend to actually here is uh
0:26:03a fact that as that already showed that that
0:26:05these biochemistry a chemistry became available
0:26:08uh i think there is a there's a lot to gain in the combination
0:26:12of chemistry tree
0:26:14specified as signal sensor specific signal sensors
0:26:17that for can look at multi spectral
0:26:19imaging as we can see here
0:26:21and and let's say
0:26:23one of the applications that we are working on a second
0:26:26a specific
0:26:27for a probe which is injected for four it to more
0:26:30a and that uh basically that broke can be used to identify
0:26:34this the lymphatic uh system and then
0:26:37it's very easy
0:26:39and some parts of
0:26:41particular search you really need to remove lymph nodes
0:26:43and this is only
0:26:45possible because of a very delicate balance between and so point
0:26:49acquisition device and
0:26:51the probes that are being used in here you can actually see
0:26:53that's removing the lymph node becomes very easy
0:26:56using these combines a
0:26:59in in finally
0:27:00uh especially uh uh yeah recently there there is immersed a lot of
0:27:06attention for personalised integrative modelling modelling and simulation
0:27:12perhaps some of you have heard of the virtual physiological human
0:27:16well especially you this has been a whole public for the past few years
0:27:20and more and more this to these uh very
0:27:24integrated models
0:27:25go from the cell like organism
0:27:27they start to percolate we into clinical practise
0:27:31uh yeah we're actually they use this this uh uh
0:27:34patient specific data
0:27:36tuned to mobile
0:27:37to the observations and from there um they can actually do predictions and simulations
0:27:43so that
0:27:44bout wraps up my uh my uh overview of uh
0:27:48what i perceive as very important
0:27:50to trends in
0:28:40i think we need them both
0:28:43uh i think of some
0:28:45i L
0:28:46some algorithms uh are so generic that they can basically
0:28:50for for
0:28:51for a lot but i think
0:28:52we also need dedicated
0:28:53times algorithms of four
0:28:55particular application
0:28:56perhaps you
0:28:57i actually my my advice there is uh for you know for those are not been working too much uh
0:29:03in the body oh area
0:29:04my advice what would be a good start working
0:29:07on the specific algorithm
0:29:10you look at the very big challenge you know the ultimate segmentation algorithm you'll have to compete against the
0:29:16all those benchmarks that that are out there so if you
0:29:19if you're working on the specific yeah a problem so you'll have people very happy at the other end
0:29:25and and and and so it will really help you also to learn
0:29:29what what the issues and of course i mean if along the way you happen to stumble on something really
0:29:36well then
0:29:37you go also for for for the more general albert but i think specific is
0:29:45there has been
0:29:45it's went let's say in the nineties to look for this uh big general super segmentation framework for instance that
0:29:51could be applied just to everything
0:29:53and now
0:29:54i see kind of more dedication again to single applications
0:29:58so that
0:29:59and and only some of those very generic
0:30:01uh segmentation framework
0:30:04they're they're still uh
0:30:06actively researched
0:30:09yeah i mean
0:30:10in most of
0:30:11areas like segmentation i mean the generated stuff is already in that
0:30:17well maybe not all music you're very lucky
0:30:20you'll find
0:30:21your generic i'll
0:30:31uh maybe uh yeah since we don't have many many uh questions
0:30:35uh so we
0:30:37i guess in our presentation we did it's just too much on the inverse problems because we thought every someone
0:30:43else would talk about
0:30:45"'cause" the you will compress and you or uh
0:30:48a those L one type of to
0:30:50as a a pretty much in the focus of of the signal processing community
0:30:55and and the a search this stuff with say
0:30:58actually those a a a a a a very hot topic in in in imaging in general so
0:31:04essentially essentially people are read is it in all the classical uh
0:31:08uh uh image reconstruction algorithms the you for "'em" C T
0:31:14in in the medical imaging area a
0:31:17the same uh for for for for different
0:31:20a could be source so that's a
0:31:22and and and not just the the traditional
0:31:25uh but that each is that also like lots of people and optics so uh designing all always new new
0:31:31a new modalities and
0:31:32then you mode that it is uh ten to work very closely with signal processing because
0:31:37traditionally for example be plenoptic optics they would just want to to
0:31:41to to do a microscope so that you see an image
0:31:44but now uh uh once you have signal processing you can do lot more like a tomography
0:31:49start a acquire measurements that you may not necessarily see but once
0:31:54the goes through a signal processing group
0:31:57then that the can can we construct images hopefully using that's measurements
0:32:02and so forth and so this all obviously they hot so if you go to medical imaging
0:32:06call you have lots of compressed sensing people the same on on the microscope beside
0:32:12so so that's so
0:32:13uh a there are yeah where the problems up pretty you well be fine
0:32:17but but then
0:32:18uh the guys also a very interested in a quantitative
0:32:23suisse showed in our example
0:32:26just to
0:32:27just to elaborate on a michael was saying is that
0:32:30but but for than them
0:32:31by your logic logical um can
0:32:34the goal has always been to do a nice image and nice image
0:32:39as you may know is
0:32:40probably don't the base one in terms of
0:32:42uh optimising devising the the rate of acquisition the rate of uh information
0:32:48it's very but the opposite
0:32:50i i actually
0:32:51if you
0:32:52a through signal processing are able to extract the meeting for information even if there is a lot of bad
0:32:57run or clutter
0:32:59to means that you can
0:33:00either improve
0:33:03the timing of your acquisition you can expose more time you're your sample
0:33:07or you can
0:33:08uh also spend more time on doing um different ranges of uh wavelength
0:33:13so you can
0:33:14have a more dense
0:33:15a field of acquisition if we were able to reduce the time that you exposing example for each of the
0:33:21the top
0:33:22in this is something that only signal processing related
0:33:25method can bring to the community so there is a whole
0:33:29range of application enough
0:33:31possible uh
0:33:32topics of priesthood for these comedians icassp
0:33:35if only you are interested to
0:33:37i mean uh the other with uh some challenging but also interesting problem
0:33:45not to mention the con the also the
0:33:48the fact that we are more immoral maybe use of the plantation there is some kind of a coming together
0:33:53of uh
0:33:54medical imaging in biological imaging and they are really merging
0:33:58in one of the key uh
0:34:01but all that is is molecular imaging which is in the the use of medical imaging techniques
0:34:06to have
0:34:07resolution which is similar to uh microscopy uh imaging
0:34:11in there there's also a lot of challenges still opened for the crib
0:34:19well there are no more question i think needs O
0:34:47i i just saw on the medical imaging side uh maybe
0:34:52not everyone will agree with me but uh it's pretty much your because uh i mean the big revolution happened
0:34:58about thirty years ago
0:35:00uh invention of summarise so i mean M R I still going very very strong and always getting better
0:35:07stronger magnets et cetera
0:35:09uh a new type of modality is drink different kind of measurements but
0:35:14it's pretty sophisticated
0:35:16now what's happening on the other hand on the biological side
0:35:20we are
0:35:21more or less experiencing this
0:35:23state of thirty years ago in medical imaging because they're all kinds of new modalities actually coming out
0:35:29almost every year
0:35:31and for example in microscopy they've been able to beat the the relay team it's by a factor of one
0:35:37hundred going
0:35:38you know like
0:35:39below what physics still
0:35:42but but by using some tricks of course and
0:35:44and they're like a novel microscopy is being developed the current the and very often
0:35:50hand in hand with with signal processing so so that's very important actually the the other thing we could even
0:35:56say it's all the medical imaging couldn't exist without
0:35:59signal processing because a mirror i the first thing it is it's the fourier transform
0:36:05and and and the
0:36:06so uh so i see that uh there there i mean there's improvement in in modality especially
0:36:13uh going out to find the resolution
0:36:15uh and and very much happening now in in the area of biology but i suppose also
0:36:21it equal so i don't know if few and this all of this small animal imaging
0:36:25yeah or uh uh so so where we getting always uh need for higher resolution source so it's very very
0:37:01oh i i i mean this is very very hard because of for example with my cross could be a
0:37:05good you use of your essence and there's lots of
0:37:08uh a uh you know like naturally fluorescent so the
0:37:13inside you used specimen so usually
0:37:15i i mean i mean this uh
0:37:17this will be a uh uh you will interact with some biologists to do the the lot labelling and and
0:37:23and i mean it's never a very simple
0:37:25but of course i mean the goal a goal of biochemistry is is finding mark is that they extremely selective
0:37:31and and also with this molecule image in principle i mean they're very good marketers soak is just
0:37:38sometimes the
0:37:39the the resolution is
0:37:40it's so great yeah so uh so we like with but um they very good mark but very poor was
0:37:45with a by you them in this since very good mark "'cause" but that terrible resolution
0:37:50and and
0:37:54K i think we have to close the session the "'cause" the next uh station is a starting and uh
0:37:58i think you all for your or of coming here and uh hopefully will see you in some more dedicated
0:38:04conference is of stations to uh imaging thank you