 from the 34th Chaos Communication Congress in Leipzig. So good morning everyone on day four of 34C3 and we're getting started right away. Here's Leander Seigl on stage and he'll be talking about International Image Interoperability Framework how cultural institutions can create an interoperable framework for digitalized works of art. Give it up for Leander. Thanks a lot. Also from my side, good morning everyone and I'm really happy that even though this is day four of the congress all of you showed up for this first talk today so thank you very much for being here and I'm really glad to have the opportunity to present this favorite technology of mine the International Image Interoperability Framework that allows us to create interoperable cultural exchange points for digitalized cultural works. About myself, I work at the University Library of Leipzig and in the Department for Digital Services and so that is my background, how I get to deal with these things on a daily basis but it's also a private passion of mine and so I have a few private side projects going on in that area. So I think it's well known and many would agree with me that digitalized images are a fundamental part of our cultural heritage for example, scriptures, paintings, manuscripts there's more access being made available through this it was restricted in earlier years but now access is provided to lots of people and so far digitalization was connected to meant that the metadata, the images themselves were kept in closed off systems there wasn't any possibility to connect them to other data and so that meant the data was also saved in a way that it was only possible to work with a certain system to present the data and so this is an example from a couple years back this was very modern back then but today this isn't what we're looking for at all and doesn't meet our demands and so how does it look like, I'm going to talk about that soon. So here an example of the situation that we were confronted with over the years and it's still that way isolated systems they have their different characters and they're accessible from repositories but there's no interoperability between them and so that is where the international image interoperability framework comes in so this build upon this build a spec and API between the repositories themselves where digitalized images and metadata is located and these APIs were connected with the presentation, the front end layer so that not even bound by certain institutions all these workspaces could be glued together by the user so it was up to the user and this came out of a community from the Stanford University Libraries, the British Library the libraries at Oxford and this community has really grown quite a lot for example the Bavarian State Library is very much involved our library here in Leipzig and one in France as well and so it's a lot of very renowned international libraries are participating in this even the Vatican Library they also provide their digitalized works of art in this interoperable framework so there's four APIs right now that are licensed under CC by license two of them are necessary to enable a presentation the image API that delivers the pixels themselves the presentation API that provides the structure and metadata for these pixels and makes them available in an interoperable way other APIs are concerned with search and authentication but I'm not going to talk about it more and so what's nice about this is that it is built on Canvas data, the data stored in adjacent format and the identifiers are usually HTTP URIs so that allows us to link different sources no matter how they're saved a few words on the presentation API the hierarchy is modeled here we have the collection that creates that bundles multiple works to a reasonable collection so researchers can create their own collections and then we have a manifest file so this manifest file means it defines a work as a series of images and then we have the canvas and that can be annotated using images or text so for example on this slide you can see that the abstract canvas object is annotated with the original image content and so this allows us to fill this canvas object with different images, with different context with different annotations and so these manifest files are structured in this way as you can see on the slide don't have to look at all the details it just gives you a rough idea of how the internal structure works and in the same way the annotations that the users themselves can create within this framework they're structured the same way and so it's possible to create annotations in my own system for objects that aren't even necessarily available within my own system but in some other institutions and again this is represented here using this triple IF we can combine images from different repositories and collect them into a single unified research environment which is a nice term I think and then save our scientific research results for the APIs from triple IF for the APIs that we use for this project for the software applications we're giving a few live examples right now we can digitally present some examples for this API and these things actually are combined from different data sources that are then shown together and then we see the function of how this application of the pictures combining together it's a dynamically structured combination of pictures here we're looking at which section of the picture is actually taken and is it rotated at all we can then call and request certain information in the server and then have certain things presented back to us it's dynamic for every application that we're using in a presentation API there's a whole series of software products for our presentation API we have a range of different applications this is an IP image server we're using now as an example we're using Apache under the hood and this is based upon the picture data there's different resolutions for the different sections of the pictures and then it is then resolved and displayed dynamically there's different types there's TIFFs that's a usually used JPEG 2000 is actually a little bit more difficult the open JPEG implementation should improve with this context the first example we have is this Papyrus from Ebers it's from Egypt it's a paper scroll that's over 3500 years old it has about 880 medical treatments written on it the roll has a length of 18 meters over 18 meters and it's now digitalized so this is now callable we can now call this to be loaded and we can look at the quality of this is very good and then we can zoom out to see the entire thing at once here 3000 pixels high gives you an idea of how large this document is and it has different functionalities on top of it so we can look at the translation that's loaded over it we can look at which medical treatments would then be prescribed for different conditions and now we're skipping over some slides this is our second example for the codecs this is the first full New Testament manuscript it was actually separated the first section was in England the second lipstick the third in Russia and also one in Egypt so this is actually a perfect use case to look at because this document is separated from each other so we need to then actually combine them to see the work in its original entirety here's a second picture of some handwriting and this was actually sitting right now in the Vatican and then there's also some in Leipzig it's in the Vatican library under our framework right now so now we're going into a demo and I'm going to show you how it's possible to use a triple IF compatible layer that you can create yourself and then use various digitalized works from various institutions I'm going to use the codecs idicus from Leipzig and I'm also going to load the manifest from the codecs Vaticanus into this layer and now I have this fewer so now it's very easy for me to have these works next to each other and using this example I can pick whatever works I want so for example if we look at the same content the start of the book of Aester and I as a layman don't know much about this but you can see that it's the same content, the same words so moving on again a representation of how this works looking at digitalized works from different repositories being loaded at the same time example number three is a demo from Project Mirador here we can see a manuscript from France and apparently this happened quite a lot an illustration was cut out of it and today this illustration is in another institution's repository and so now using this triple IF model we can combine different digitalized works on the same canvas so now we can recreate the original impression using these digital tools and since this is such a nice system I can show you how annotations work now so we can define certain area and then enter our text important notice and depending on how the fewer system is configured this annotation gets saved locally but it will be possible in the future to have a work environment for researchers to deposit their annotations in some central system here another look into the manifest file here we can see that the URL is loaded from different image servers and those two images get combined on the canvas layer another example we have this technology tested in the real world at the university library of Leipzig there's a handwriting course every year and so we provided a work environment for the participants in which they could use our own manuscripts but also manuscripts from other institutions and annotate these manuscripts working together and so document their research results together and this was provided by the Stanford University Library using this Mirador system that I've talked about already and as I've said in the beginning I want to talk about a little private project of mine using wiki data and the SparkQL endpoint I loaded images of caps onto my own server changed the format to IIIF and made it available for the public and searchable and so now in this system you can pick the digital works and now I can also show you the drag and drop feature that is recommended for implementing the discovery system with presentation systems and through this system you can also reach IIIF manifest files to use these images in different contexts and so another interesting fact a lot of institutions have used the IIIF system to provide digital content online for example the Metropolitan Museum of Art in New York another institution in France and also one in Switzerland and they have this project called E-Codesies and you can look at the PDF presentation later and get the URLs from there archive.org also supports the standard by now so this to give you an example of huge data pools that you can tap to fill your presentations or your projects so these IIIF APIs are constantly being developed further there's going to be a switch of the data models for different annotation possibilities and we also want to support video data looking at the institutional side of things the University Library of Leipzig passed the Open Digitalization Policy so most of the digitalized works we want to publish them under CC0 or CCPD mark licenses and also IIIF compatible and so if you go on this URL you can look into our digitalization workspace and see what is being digitalized live one more remark we're going to have a hackathon soon in Leipzig where we will also host a IIIF workshop thank you very much for your attention if you want to learn more about IIIF then you can check that out at IIIF.io and here's my contact information and if there's any questions we still have some minutes left applause thank you very much we have five minutes for Q&A if you have any questions ask a signal angel is the internet smart and has questions no questions from the internet let's start with the microphone good morning thank you very much for the presentation do you have any questions about inventory data banks my question then is about the database what text is there exactly so this product I can't say anything about that but I believe IIIF is a well it's just a description of a standard framework and so systems would have to provide access points to realize the standard my question is there's now a standard are there alternative standards I mean the big advantage here is that everything you can alter it you can apply it but what about if Berlin uses a different standard for instance and then how do you then in common apply all of these different standards yeah there's an older standard based on XML files in which data is supposed to be delivered as whole images but the technological standard we provide layered images and that's really we're one of a kind there any more questions no everybody's still tired all right so thank you very much thank you for listening to the interoperability