 I am again here with a new topic called multimedia database, so today we are doing an extensive use of multimedia content for e-learning, like we have on a YouTube in that we can watch the videos, we can listen the songs, we can watch the movies online. So we are, so we uses the multimedia content everywhere, so before going into detail about what is multimedia database, the knowledge associated with multimedia database has to be made directly accessible and but the searching or accessing the multimedia content is still a challenging task and therefore we uses the multimedia database. So what is the multimedia database? The multimedia database is the database that host one or more primary media file type. Now what are the primary media file type? It can be a .txt that is a document or hand written document, it can be a audio with an extension .mp3, it can be image with an extension .jpg, .bmp, .png or it can be a video like AVI video, WMV, .mkv, .mpeg. Now what is the multimedia data? Multimedia data is loosely fall into two main categories. The first one is a static media which is a time independent because it is a static, it does not have any relationship with the time that is the images and hand written documents and second one is the dynamic media. Every dynamic media has a certain time limit, the video plays for a certain time, the audio plays for a certain time. So it has a relationship with the time and therefore it is a time dependent. So dynamic media contains videos or a sound byte and these all the primary media files are stored in a binary strings of 0s and 1s and are encoded according to their file types. Now how the multimedia data is different than the normal data like string or any other data? So multimedia data is large and because video has more video, video takes more memory in a computer and since it is a large it affects the storage and since it is a large the retrieval and the transmission of multimedia file becomes hard or it will take more time than the other data. In case of a video and audio databases time to retrieve the information is more than the normal database. As the size of the data is large we need a certain special data structure for extraction and indexing. So which data structure we are going to use? We are going to use the enhanced multimedia meta object. In short we can say it as EMO. EMO is an Enrich knowledge unit of multimedia content which comprises the media aspect, semantic aspect and functional aspect and it also supports the version aspect of each multimedia object. Now what is the media aspect? The media aspect contains the multimedia files like video, audio, images that is in the form of binary, binary strings. We can store the video or images in the binary format. The semantic aspect we have in a relationship with video and its slides like one video has many slides. One slide has a start time, end time, these type of relationship contained in a semantic aspect. The third one is the functional aspect. Whatever be the action involved on a multimedia object that contains in a functional aspect. Now EMO is a formal entity which comprises the four parts, logical media part. Multimedia part means, in short we can say it as a LMP, LMP means the multimedia file that is the audio, images, ontology object means every video has some attributes like its author or creator like that. Association, association can be that one video uses these many slides, that kind of relationship means association and EMO nodes contains all these logical ontology and association. It forms one EMO node, we will see this using a CS101 lecture. This is a one EMO entity in that we have a lecture video see here, in that we have a lecture video, lecture video uses the PPT, it has a lecture slides and the lecture video has a LMP this through the URI. Now what is the ontology here? The lecture video has a topic, topic has a value introduction to programming, these are the attribute topic, this LMP is a kind of video. Now this lecture video has a creator attribute and which has a value a club here and the speaker of this video is Professor D. B. Pater. Now this is for and one more the format here, the format is mp4, that is because of EMO we can get the smallest information like the format ffmpeg, it is a media codec, audio codec is HAL and it is a full because we are using the full video, therefore it is a full. Now the lecture slides, the lecture slides also has a author, that is the Professor D. B. Pater, it has a format PPT, number of slides 29th. Now each slide has a timing start time and end time, that is first slide will start from 0 to 1 minute 4 seconds, as soon as the topic will change the next slide will play on 1 minute 4 seconds to 2 minute 35 seconds, like that all the slides will play according to their topic. Now this is the EMO representation for CS11 lecture, now how can we implement the EMO in our database? EMO is a nothing but a model for giving the information, you do not need to learn the EMO, but it is just a structure which it is just a pattern which designs the database as that we can use or we can access easily or we can retrieve or insert the multimedia data from the database easily. So for implementation we have to realize on a distributed infrastructure, we have to, we need to have a tables in a distributed manner so that we can get the smallest information from it. It is intended to enable the sharing and collaborative authoring of multimedia content and we are going to use the any of the technologies for creating the database like PostgreSQL or MySQL. We have to implement this EMO, so as it becomes a platform independent, I am requesting Mr. Nagesh will talk to you better on, this is the EMO model that is Enhanced Multimedia MetaObjects, actually there is a PhD thesis in 2003 something like that, this model actually gives you a power of representing information. Now previously we had just normal databases, now for databases you have to design certain schemas, certain schemas are there like you want to say employee, employee as a table, then employee code, then employee name, so you have to know the ontology well in advance, ontology is something like say Eskimo, Eskimo says this is white snow, so Eskimo may even say this is a black snow, green snow but for me it is a white snow, now if I can say white snow for me, for Eskimo I will have to tell him more information to say that this is a black snow, for me it may be white snow, for me snow is white, for you also may be, snow is white or black, you have not seen green snow, no? Green snow, green snow, brown snow, different different, so it is like you know you are trying to match, you know what are the meaning, the meaning of this you are trying to match with something else, so two different communities may have different meaning for the same thing or you want to come to an agreement, one agreement that is ontology, so why do we need ontology, I am not sure why do we not need ontology but it is always useful because when we design database is ontology is right inside the database, right inside the database, you do not see it, you only know the schema of it, so this is little mathematical what I think is understandable, it is only set theory nothing else, so set theory I mean we must have had introduction set theory right, quite easy, so now this model actually is based on a 13 tuple model, 13 tuple, so what it says is that you can represent any information in this 13 tuple, any information, you want to represent time, you want to represent spatial data like GIS, geographical information system or if you want to represent even time independent things, 4 dimension, 5 dimension, 10 dimension, so any structure you can represent, so you can see that W an entity is a 13 tuple, now every tuple will have a unique identifier just to identify that it is unique because in database is you need to know what is unique, otherwise you will not be able to locate it where it is, so certain unique identifier is given to that particular tuple and then you have a name of that W what else is that kind, so there are 4 different kinds of objects, so one is LMP as he had already said LMP, LMP logical media part, second one is your anthology, third one is association, fourth one is MO, now these 4 they are all different, LMP actually denotes the actual resource that you are having, now you want to do something with that resource, the file, media, this that whatever the physical proper resource, anthology, anthology gives you relationship between different resources, you want to say CS 101 lecture in IIT Bombay and you want to say CS 101 lecture in Stanford University, how do you say they are the same like a white snow and a green snow, there is a slight difference but you want to say that they are same because there is some programming involved in that lecture in those lectures, so that is anthology, third one is associations, you want to associate after this lecture, this lecture, second lecture, third lecture, fourth lecture, there is there is a temporal activity there or you want to represent much of the time related things or much of the you know versioning mechanism like I took a lecture from Stanford University and I made my own lecture but then I modified say one fourth of it, but remaining three fourth is the same, so then I can keep a track saying that this is the original one, this is the modified one, I put a timestamp, timestamp, so I know from where I have copied, so that is the association and MO, MO itself is in a hierarchical fashion, so it keeps on encapsulating everything, so if I go to a figure, this is about Dracula movies actually, all different Dracula movies, so this is Dracula movie, I think this is the original one, this was created from this, so there is a predecessor and a successor, the whole thing is an MO, I can actually put the whole thing in an MO, so it is all hierarchical, so I have predecessor, successor, I have predecessor, version 1, version 3, version 2 is here, version 3 is there like this, this is the way you would actually like to represent, it is a complete information that you would like to store in a database, so how do you do that? If you had to design a system without considering this, this model, this is a model and the design itself will be different, this is the model, if you had to consider doing a database design without considering this, then the database might be very huge, very huge, if you heard of temporal databases, the special databases, we take care of only time, databases we take care of only time, so they are temporal databases, they are special databases, they are not normal databases, you must have heard of post GIS or post-grace SQL, MySQL, MySQL will not consider time, post-grace SQL will not consider time, post GIS will consider special information like related to GIS, geographical information, so there are time, time DBs, we take care of only time, so different types of database are there, for special, special use, now what we want, we want one, one which can have everything, so if the database is not taking care, you have to take care at the model level and what, what is that they are trying to do is that they want to take this model and put it down, break it in the form of a schema down, so that the redundancy is removed, because in fact there is a redundancy, if I, I will show some examples, so this is an example, logical media part, LMP, that is a unique ID number and based on that 13 tuple, so if anything that you want to represent, you put a 13 tuple at the model level, it is a model level, then the name, then it says it is an LMP, logical media part kind, LMP there, then those are null and null, I mean those are null and null and in many places you will find, now you see there are many places where it is 555555, it means nothing is there, so if you want to represent even a small thing like somebody has a researcher, which is an ontological object, whatever is O, O is ontological object, so I want to represent ontological object, I want to represent association, association object, I have to represent logical media part as an object, it is just a representation, what you see is a representation, so I have ontology, ontology, this is the ID and then everywhere null, null, null, null, null, null, null, so if you do like this in the database, then it will take long, long, long time, I mean lot of time to actually retrieve particular information, so now what they have done is that they have converted these type of data which is coming, so they have converted into a schema, a brief schema is available, they have some problems also, now the problem is some of the information to fetch is actually taking time, even after doing that it is taking more time, so probably your job would be to improve that, that is for a, to improve the performance of the system because nobody would like to wait for 5 hours, if you press coming after 5 hours, nobody would like to wait, are there any questions, so now you see a bigger MO, this is one, inside one MO is there, this is MO, model, model, this is big one, so inspire is an association, whatever is, this is association, so you have remake, vampire, Dracula, so they are inspired by, so you have to create those associations, you have to create those ontological objects and you have to create, I mean not you have to create, I mean you should allow a person like me who does not know anything, I mean a layman to create something like this, so you have to, you should be able to map it, that this is an association object and then it goes inside a database and sits properly, like for example if you have a form and then you say name, roll number, this, this, this, this, where it should go, you have to identify or somebody is putting a lecture video, somebody is putting a slide, where it will go, so those connections you have to identify and provide that path, so this is mostly an occurring tool, you will not go and go inside the database and put labels and things like that, so you have to provide something which automatically takes and puts it there and how you connect, you want to show that they are inspired, you create an object inspired and then you may want to search, is there another word in the dictionary which is same as inspired because ontology and if it is same, then you may not be, you should not be allowed to create another object by that name, so ontology will actually bound you, I mean the rights for having ontology in the system will be totally with the people who will be creating that system, should I allow that this ontology, should I allow this ontology, which ontology is better, author is better, writer is better, movie is better or film is better, so you have to choose before creating an ontology, so you will not create, you will not choose that, but if you take that systematic examples while designing this system, then you should provide such type of facility for choosing or changing, is it understandable, so all are, I mean all the examples are here, I had asked him to take 3.2 and 3.2 is here, so you can circulate 3.2 to everybody, just simple example, you are not very difficult, just everywhere there is 0 0 0 0, so you can understand how simple it is, in brief I have covered MO, so now in detail if you want to know what is MO, go through those examples, create your own examples like that lecture CS 101 is something, just to understand and then look at the schema, I think there would be people who can help you in knowing whether the schema is proper or not or it confines to this model, if it does not confine to this model then the schema will have to be changed, I had looked at the schema long back but I do not remember, so probably if I get a chance I will look into that schema, so but questions are important, in the morning I said questions are very important, two way communication is better, what if you had a mobile which would allow you to use as a walkie talkie, one way communication, so this is around 139 pages, so do not read 139 pages, read what is appropriate and also there is a algebra for this, what I what you saw is set theory but there is an algebra for it, go next, next, next, those have been implemented that is called MMA algebra, so you know what is an algebra, like in what algebra you have in database, relational algebra, same but this is object based where you want to say next or I go to the next or I want to iterate or I want to do something, get me that particular object, so algebra will tell you what to do and there are lot of examples given how to use that algebra properly, so maybe you can look at that algebra also and if you can build, see I believe that is slow because the algebra is not implemented properly, otherwise it would have been faster, so if you can look at that algebra, create that algebraic querying, then it will be faster, that layer is required, otherwise your databases would have never been faster, the relational algebra did everything, you know, really it would have never been faster, indexes and all those things have really helped databases, query algebra is there, try to query, you can look at this, once you understand this model it is easy, you do algebra no, 9th standard, 10th, model you may not do at 7th standard, 8th standard model is some, that is more important, model and now how you do it inside, if you read it is quite easy to understand, not very difficult and if you have any problem probably you can contact me but maybe through them or whatever or I do not know what is the mechanism, what is the process, so whatever is the process I will be there available.