<TITLE: Gaze Path Research Seminar
ACADEMIC DOMAIN: technology
DISCIPLINE: information technology
EVENT TYPE: seminar presentation
FILE ID: USEMP04A
NOTES: continued in USEMD090, seminar also includes presentations USEMP04D-E

RECORDING DURATION: 19 min 25 sec

RECORDING DATE: 4.3.2004

NUMBER OF PARTICIPANTS: circa 30

NUMBER OF SPEAKERS: 4

S1: NATIVE-SPEAKER STATUS: Finnish; ACADEMIC ROLE: senior staff; GENDER: male; AGE: 51-over

NS2: NATIVE-SPEAKER STATUS: English (UK), French; ACADEMIC ROLE: senior staff; GENDER: male; AGE: 51-over

S3: NATIVE-SPEAKER STATUS: Croatian; ACADEMIC ROLE: senior staff; GENDER: male; AGE: 31-50

S7: NATIVE-SPEAKER STATUS: German; ACADEMIC ROLE: senior staff; GENDER: male; AGE: 51-over

SS: several simultaneous speakers>


<S1> <USES POWERPOINT> okay good afternoon everybody i think we should get started er , welcome to this seminar on gaze path research , we're expecting close to 40 people and i think there's something like between 25 and 30 of you so some may be arriving late but nevertheless i think we should get started er good to see you so many people from outside the university as as well especially when we have what we think is a very interesting programme er you can get a copy of the programme when you walk out from this er auditorium during the break and an er brief abstract of the of the presentations the same s- things which are on the web with the handouts , on the desk there by the door , so er we'll start with a brief introduction er by <NAME S4> from university of art and design helsinki and myself will tell about er why we have organised this event sort of put this into perspective and then we are very glad to have three distinguished speakers here professor <NAME NS2> from camberwell college of arts er in london institute to (become) the university of arts london right </S1>
<NS2> that's right @@ </NS2>
<S1> welcome er after after his presentation we'll all move to the new part of this complex pinni B and we have coffee reserved tables for us in the cafeteria in that building the cafeteria is called minerva and we continue in then lecture room er B three one one six if i remember correctly which is very very close to the cafeteria just one one floor up from the caf , and after that we'll hear two more presentations er first professor <NAME S3> from university of art in philadelphia and and er then i think we'll even though it's not on the programme we'll have a five minute break between at least and then professor <NAME S7> from er <FOREIGN> fachhochschule gelsenkirchen </FOREIGN> correct </S1>
<S7> yes </S7>
<S1> so a warm welcome to our all our guests and the theme of all these presentation is somehow related to using eye gaze in an artistic or design context so you see that there are these two sort of themes which er i think run through all the presentations and it's also er sort of , reason why why this seminar has been organised it's been organised in some sense by many bodies so i think the first of this was a project called affectionately called <FOREIGN> mutu </FOREIGN> in finnish it has two names it's called POD in english and this is a project we started last summer er it's funded by tekes the national technology agency and it it's belongs to the <FOREIGN> muoto </FOREIGN> 2005 programme design or shape 2005 actually but it's a design research design technology oriented programme and this is a cooperation er project . er which is being carried out jointly by our university especially the research unit on new computer human interaction and and there the gaze research group on the one hand and then university of art and design helsinki school of industrial design on the other hand so we have four researchers employed in that project currently , so this was a first cooperation activity and then the academy of finland launched a sort of companion programme research programme called industrial design and there we are lucky enough to have actually two projects one is er called gaze path research supporting designers and users it's very closely related to the first one perceptional design so we are gonna study the perceptional design using gaze path analysis techniques and actually even though i've just drawn one arrow there this is also a cooperation project between both universities so <NAME S8> from university of art and design helsinki is in charge of er part of it and i'm in charge of the other part and then <NAME S4> also has a pro- project in that programme domestication of design so all of these are working in sort of a er , er together on on similar themes but from different points of view and then the e-tampere project a- and <NAME> has been also active both in organising this seminar and also in getting all of these project- -s projects started so it was a very good resource for that purpose . so so we have these two actors here our main interest came from the from the b- er gaze study research and of course the group of people at the university of art and design are most interested in design and er to put this er project sort of into context i'm planning to er give a brief overview of what kind of gaze related research we do here in this university also introduce er some of the people doing that research so that if you are particularly interested in in gaze research you know who to get in touch with during the breaks or after the seminar , we've been doing this for many years the first device was er acquired around 1995 an ASL remote controlled eye-tracker for those of you who know this technology er and these manufacturers then we got i think around 2000 an S-M-I eyelink or distributed by S-M-I a german company er head mounted tracker more accurate also more inconvenient in that it has to be very tightly strapped around your head er and two years ago in a european project we got er other remote controlled tracker eyeview X which is er i guess the most convenient one of those which we have currently and then within the next two months we hope to get yet another one two B technologies eye tracker and if you if you see how the technology has developed this w- this was er state of the art about ten years ago and that's only only very small part of it this is what's er visible to the person who is who's being tracked then there's a big box huge er old-fashioned piece of equipment on the other side used by the the person who operates that system and and it has all been st- er shrunken into into this size so it's actually embedded into the monitor there's no separate electronics needed anymore so it's becoming more and more convenient to use and fortunately it's also getting cheaper er not as fast as we would hope for but but anyway . so we've got a lot lots of these erm and of course there must be a reason why we are buying new trackers every two years or so , other than that that they're getting better we're doing using them for many different studies and i'll give you a brief overview of how eye tracking can be used the first er and i think this is er oldest of this field which is command based gaze interaction er er this is a area where the most most of the applications which have been built exist and they're typically used by er users with the severe disabilities typically motor impairments or users who can more or less use only their eyes or only their eyes and use their eyes for communication so eye-typing systems which you can see here er er soft er keyboard is played on the on the screen and letters activated by gazing at them is a classical application , it sounds like a very slow er way of communicating and of course it is compared to our normal communication but for those people for whom this is the only option it it it's it's really er opens a new world for them there are these several books written using this technology so er for instance people who suffer from the locked-in syndrome like like er philippe vigand or or people who suffer from ALS lou gehrig's disease have written er books about the the development of of their gaze and and that has been done using eye tracking systems , what we are doing here is is try to er sort of create science of this er in this context so lots of these systems have been built but they have not been studied a- and all the all the developers have their own view of what works and what does not work and why but there are hardly any publications so what we try to do is to carry out controlled experiments to really understand the details how long should the dwell time be how should the interaction be er be designed so that it's smooth and another goal here is to design new widgets which are specific for eye gaze not just use the soft keyboard on the screen but have widgets which are useful for editing it's very hard to position a (character) inside two letters in the text if you just try to use pure eye tracking you need some some help some new widgets to make that possible , er the people who are er working on this er . field have have joined a network recently and proposed er er . network of excellence to the european union and and it's still under review getting actually into the pre-negotiation phase so we are still optimistic that that will go through and i think that will sort of op- increase a lot er er cooperation between these research units and also the cooperation between the user bodies user communities and the researchers <NAME> er is a graduate student in in the U-C-I-T graduate school and this is her PhD thesis topic so if this is of interest to you get in touch with her , she's smiling but others will be smiling too </S1>
<SS> @@ </SS>
<S1> er also we are developing new kinds of interaction techniques er not only in this context but in general windows contexts er <NAME> he is a post-doc and <NAME> er there is is a PhD student doing research o- on these aspects , okay another field of research is is attentive interfaces sort of non-command interfaces where you don't explicitly use your eyes to give commands it's not a very natural thing to use your eyes for but eyes are good in in indicating where your attention is and if the software could react to that perhaps the interaction could become more smooth than it's currently is this has many names attentive interfaces non-command user interfaces proactive interfaces all mean more or less a similar thing and i think <NAME S3> for instance is going to talk pretty much about this kind of issues i believe </S1>
<S3> yeah i believe also </S3>
<S1> good </S1>
<SS> @@ </SS>
<S3> help </S3>
<S1> the thing which we have done here is is d- we have built er er environment for reading documents which are written in in a foreign language and and the design is such that we have on the screen we have the the text which is which the user reads and then we have a separate frame for dictionary entries and the idea is that the user just reads the text in the normal fashion but when he or she encounters an uncommon unknown word probably the the gaze will get stuck and and she will stare at the word for longer than than usual and in that case sort of a er gloss appears on the screen pops up automatically and gives a translation in one of the languages supported by this system , er if this is not enough then the user can sort of switch into this command mode and deliberately look at the dictionary frame on the right er and and he will get the full dictionary entry for that word so that gives more information . er the hope is to make this smooth so that the user hasn't even noticed that that something is happening but that er the interaction is is really natural that turns out to be a big challenge for many reasons the least of which is not the inaccuracy of eye movement tracking so it's really hard to to say for sure which word the the a reader is is reading when when the text is displayed on the screen the prototype for this one was developed in a again in a european cooperation project and <NAME> is is doing continuing this work as her PhD project . thirdly er we have been developing and using eye movement analysis tools so this has some early pictures of our products which have been tracked in this current design-oriented project this is a typical way of of showing how the gaze behaves the radius of the circle denotes the duration of the fixation how long the users stay at at this point at this point of time and then the lines to present (arcs) which connect the fixations so if you do this er interactively one by one st- one by one you can really see how the gaze moves when observing this object many times you need other kinds of of visualisations for instance this kind of an overview er intensity map which just gives you an overview of where on which spots the user's gaze landed which what points caught the user's attraction you can do thi- compare these for various users or here you can have the data for two users and you can see that they er looked at some of the same spots like here some of the different spots as as well this user looked here more than more than this one so you need different visualisations for different purposes and and er that we have been developing again in the project where <NAME> is working and his tools have been used in many of our projects one of them is is being carried out by <NAME> who's going to study er the layout of web search engine result listings so whether bolding helps if then it doesn't help why do users really look at bolded words or do they look at words dif- er er rendered in a different colour and so on but for that we really need need high accuracy eye-tracking and we hope to get that with the new tracker . and then finally this current project so this this is a background which led us into this field into proposing this er project of using eye-tracking in in in design we er really wanna understand want to understand more in more detail how the eye gaze behaves the way it does what attracts attention and this is er s- so- sort of promising area for studying that @@ that question er we are going to do that er using a series of empirical studies where we vary different variables and and study different conditions designers versus consumers attitudes of consumers will probably be taken into account as <NAME> pointed out yesterday the tasks wha- how the users are instructed to to er view the pictures the fidelity and type of the artefact i'll show a picture of that er from the technological point of view this poses some interesting questions er eye-tracking in 3D hasn't been done much typically eye-tracking is used in a in a monitor environment where the distance between the viewer and the thing which is been viewed viewed is fixed and is not too large we would like to view eye-tracking with things which the user takes in his hand can freely move around in the space , er that hasn't been been done much so far tracking of large objects is another other challenge some of the design objects which we are going to study are are bigger than you can you can er realistically display on the screen or or hold in your hand . er so this illustrates some of the variables the representation of fidelity from IBS to products from 2D pictures to 3D models on the screen to real objects and and this is a- already this is a large base to to st- er investigate then you've had if you add the other dimensions designer versus consumer and so on there are really many many issues to that er the researchers working in this project are <NAME> <NAME> and <NAME> all of them present here and er , hopefully in the near future you will hear more about the results which they obtain in this project but not today , this was mo- more or less the background of of our interest in this field and now <NAME S4> will continue on why design researchers are interested in this area </S1>
