A Report On Multimedia Database Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Multimedia database is widely used to store the large amount of data. MPEG-7 is a metadata standard which describes the details about image, audio, video, graphics, etc. In this paper mainly concentrates with the content based retrieval while using MPEG-7 from multimedia database. And it also describes the MPEG-7 methodologies and its issues. MPEG-7 describes descriptors and description schemes, and these two relationships defined in the language DDL (Description Definition Language). It mainly deals with integration of MPEG-7 into Multimedia database using Data models, query language, access methods and query optimization.


Multimedia data used to store in the multimedia database by using metadata. Metadata describes the information of the specified multimedia data or object or file. Multimedia data contains image, audio, video, graphics, 3D model, etc.

Media data:

It constitutes the actual data as images, audio, video that is digitized and captured, processes, stored and compressed.

Media format data:

It contains the information concerning to the format of the data about media after it goes through the acquisition, processing, and encoding phases. And it comprises of data such as frame rate, encoding scheme, sampling rate, resolution, etc.

Media keyword data

It contains the keyword descriptions about the keyword matching and it perfectly matching to the media data generation. it consider the date, time, and recording place , who is recorded, recorded scene, etc . This informative is also known as content descriptive data.

Media feature data

It defined the features obtained from the media data. The media contents are characterized by this feature. It particularly specifies the color of the image, texture information about the different objects and shapes descriptive to describe the multimedia data. It is also called as content dependent data.


MPEG has introduced the video and audio compression standards such as MPEG1, MPEG2 AND MPEG4. Mpeg7 is a new metadata standard which introduced by MOVING PICTURE EXPERT GROUP in the year 2002. It officially called as MULTIMEDIA CONTENT DESCRIPTION INTERFACE. It described the features of multimedia content and it also determines the audio-visual descriptors that can used to depict the different kind of multimedia information. It has number of system tools such as Descriptors, Description schemes (DS), DDL (Description Definition Language) to described how metadata used to describe the content of multimedia data. It can accept to support much different kind of natural languages. Multimedia data not only consider the image, audio and video but it also contains 3D models, speech, graphics and details on how these components are integrated in a presentation of multimedia. It affords effective content based listing and retrieval of multimedia system. It exposits the structural, semantic and low level features of defined multimedia file.

Descriptors and Description schemes

It represents the syntax and semantics of feature representation of mpeg-7. It provides standardised audio visual descriptors. Such as color descriptors, text descriptors, shape, etc.

It defines the structures and semantics of components and its relationship between the descriptors and description schemes.

DDL (Description Definition Language)

Description definition language grants the new type of mpeg7 descriptors and description schemes. The existing description schemes can be modified and extended using DDL. The DDL schema is converted in XML and used markups to determine the document structure and embed it. DTD (Document Type Definition) defines the explanation of reasonable structure of the document. XML is also used as same as HTML user definable tags. DDL needs particular extensions to XML schema language to satisfy all the MPEG-7 requirements. Extension such as MPEG and W3C are discussed with their links between them. MPEG-7 DDL used the base language called W3C XML schema language and it add to determined MPEG-7 mechanisms like array and matrix data types. The DDL parser is used to analyse the description validity with using several rules. The parse is mainly used to initialise the schema and delivers the description of errors report. Presently there are only several schema parser is available in MPEG-7 but XML schema parser(XSV, XMLspy, Xercer) is used to validate the description schemes and descriptions.


  • Content based retrieval
  • Integration of MPEG-7 into multimedia database


In this section clearly explains the how MPEG 7 interacts with multimedia database systems which used as data model, query language, access methods and query optimization.

Data model

Data model is a essential component to retrieving and managing the multimedia data from the database. Data model is dividing into two types of approaches such as structured storage approach and unstructured storage approach to describe the storing operations and it can support retrieval techniques. MPEG-7 trusts on XML-schema, mapping strategies for XML have to be considered by an equivalent database data model. In this model consider the mapping problem in mpeg7. Data model which defined as high level multimedia database system based on MPEG-7. It also consider mpeg-7 documents to be used as inserting, deleting, updating facilities to integrated into the database system.

Query language

Multimedia database systems broaden the object relationa SQL-99 for the use of conceptual multimedia data model in SQL/MM.SQL/MM data model covers the descriptions of multimedia syntactical part. But it allows the process image decomposing for describing the meaningful content semantically. MOQL (multimedia query language), it extends the object query language used to adding the content based multimedia data retrieval properties such as spatial, presentation and temporal. Mpeg-7 as a data model, it is mainly used to face the enhancements of the query language to searching the multimedia object. In this context, it also defines the integration of operations that can delivers the XML output has to be clearly analysed as well. It should correspond the both (MPEG-7 descriptions) import and output formats to each other. It refers the retrieval of query result as a document of MPEG-7, and it is used to combine the SQL/XML elements with multimedia query language. It verified the resulting the XML document conforms to the MPEG-7 XML schema. It requires the enhancement of query processing for MPEG-7 conformance type checking.

Access methods

In modern database systems are mostly used to efficient searching and retrieving concept with help of Indexing. Innately, only limited number of integrated access methods (B-tree, hashing facilities) is used in most database systems. In the process, integration of such access methods using similarity searches and query types to support the retrieval of multimedia content. In this section content based retrieval has been used indexing and can support low level features. Although Semantic indexing and querying are defined by semantic descriptors in MPEG-7.

Query optimization

In multimedia databases, low level feature defined the queries to execute some similarity operations such as nearest-neighbour operation or range operation .query optimizer is used to extend and improve these operations activity. Query optimizer is classified into three approaches. There are selectivity, cost model and operator. Cost models concentrate on calculating the cost of index structures for nearest-neighbour and range neighbour searches and it used index tree in k-NN(k-nearest neighbour). In selectivity process represents exact low and mid dimensional data with non-uniform distribution.


Content Based Retrieval (CBR) has described the following sentence:”It is the way to retrieve content, based on what is in the content”. It mainly focuses to define the multimedia content retrieval done by using keyword matching. Multimedia textual information represents the title, author, description of content, creation time, disclaimer, copyright, details of software used, warning, comments and source. It has two types of features such as low level and high level features. Low level features specify the color, shape or texture. High level features are classified into two classes (level1 and level 2) based on its level of abstraction. Level 2 represents the particular image or multimedia object and also denotes specific interest of region about the object.

Level 3 specifies the semantic circumstances of acting objects or represents the scene. It also specifies the extraction algorithm of level 2.

  • For example, a Portable Network Graphics (PNG) image file can contain textual information such as title, author, description, copyright, creation time, software used, disclaimer, warning, source and comments.


Color is one of the main low level features for retrieval of multimedia content using color histogram which depicts the proportions of pixels to find out the image. For retrieval time user text will compare with stored images and it displays all related images. The main advantage of a color histogram is its lack of information position.


Shape is also most useful to recognize and retrieve the multimedia objects by using primitive level requirements. It represents rotating, translation, scaling to find out the objects. Shape representation can be classified into two types of algorithm such as Boundary based and Region based shape representation algorithm. Boundary based, it defines the outer line or boundary line shape of the objects. Region based, it used to determine the seal shape region of the objects. Fourier descriptor and moment invariants, these are two examples of shape representation algorithms.


It is mainly used for retrieve the several types of images or objects based on between same object representations. It depicts texture representation such as coarseness, directionality, contrast and regularity, or periodicity, directionality and randomness. In Texture representation the most widely used filter is Gabor filter.


Content based retrieval is most useful for retrieving the image, audio and video from multimedia database in World Wide Web. In MPEG-7 metadata, it defines the information about the multimedia data and easy to retrieve the content using keyword matching concept.


Descriptors is most useful for describe the information perfectly to store in the database. It classifies two types. There are visual descriptors and audio descriptors.

Audio descriptors

It represents the description tools describes searching the voices, similar envelopes and frequencies of sample voice against a voice databases. It classified into several types such as Basic descriptors and Basic spectral descriptors.

Visual descriptors

It contains several types of descriptors. There are Color descriptor, Texture descriptor, Shape descriptor, motion descriptor.

Color descriptor

It depicts the information about color and how it is used to retrieve the content. It consists of seven descriptors in it. There are Color space, Color quantization, Dominant color descriptor, scalable color, color layout, color structure description, GOF/GOP color (Group of frame/Group of pictures).

Shape descriptors

This descriptor defines the boundary line or edge line of the object shape. This describes about object retrieval using defined description about shape of the object. It divided into three types. There are region based shape descriptor, contour based shape descriptor, 3d spectrum descriptor.

Texture descriptor

In this descriptor clearly explains the content about text based store and retrieval from the database. It classifies into three types such as Homogeneous texture, texture browsing, edge histogram.

Motion descriptor

It determines the video segmentation, accuracy of the video, estimation, time and place, frame segmentation, etc. There are several descriptors to describe motion retrieval from the database. Such as camera motion, motion trajectory, parametric motion descriptors, motion activity. ( Eidenberger, 2004).


In this paper, I have fully concentrated with how retrieve the multimedia data from database integrates with multimedia database. And it also annotates the drawbacks, advantages and what problem rise while using time and how to solve the problem. These questions are answered which I have used to compare the journals and several author discussions, approaches on their papers. I have chosen the paper “MPEG-7 based image retrieval from world wide web” author Agarwal Etal. In this paper, the author used three layered (integration, intermediate and distributed sources) integration architecture to retrieve the image and retrieval from database. There is no need of semantic information to store the images on database. For integration of schema technique we need to face many challenges. There is a declaration in the standard for MPEG-7 before the image has been stored in metadata format. It would be a great mistake for the metadata to be available on MPEG-7. Different researchers suggested many low level features, even though there is an inclusion of MPEG-7. There will be a new feature for this release in the future and it should contain in the basis of system image-retrieval.

I have chosen another paper “Image indexing and retrieval System using MPEG-7” author Kyung Lee to compare with previous defined paper. In this paper, the author concentrates with image indexing and retrieval using two descriptors such as homogeneous texture descriptor and color histogram descriptor to describe it. The paper shows metadata to find the images in database through text based retrieval. Text based retrieval is best one because it is easy way to searching and find the object. Image, audio and video file format based searching and retrieval is difficult but text is easy to stored information about the object and using keyword matching to find and retrieve the object.

And I have took another two papers to analyse the audio and video retrieval using MPEG-7.the first one is “A Video Browsing Application Based on Visual MPEG-7 Descriptors and Self-Organising Maps” author is Horst Eidenberger and second one is “AN MPEG-7 DATABASE SYSTEM AND APPLICATION FOR CONTENT-BASED MANAGEMENT AND RETRIEVAL OF MUSIC” author is Otto Wust. The second paper mainly deals with how music retrieve using an application called Sound Palette. The author depicts the information about the application and its musical retrieval. It concentrated to use as data modelling to implement the database to do the retrieval. I have referred sound palette is how undergoing and how it to do the music retrieval process. In the first paper, the author concentrates to describe the visual descriptors and its characteristics. The author annotates “The major advantage of compressed data-based approaches is that they require less computation power than approaches working on the uncompressed domain. Methods for detection of effects are usually based on feature-based approaches”.

Finally I have took the paper “MPEG-7 meets Multimedia Database systems” author is Mario Dollar. The author described in this paper how MPEG-7 integrates with Multimedia Database Systems. In this paper mainly deals with integration, it covers data model, query language, access methods and query optimization. These all defined and delivered several solutions and proposal to storing and retrieving the documents of MPEG-7. It delivered every solution enforces different combinations of used retrieval operators and query languages. In future Query language will must develop the input and output format of MPEG-7 queries. In this paper author defined the unresolved problem of index structures using Data model because it has limited availability of index structures.


In conclusion, I have discussed four journals to explain the content based retrieval which describes the retrieval of image, audio, and video. I have learned to get knowledge about how to retrieval the multimedia content and find the solution of some problems and unresolved problem also. And how MPEG-7 integrates into multimedia database. The biggest issue in multimedia database is the retrieval of data from the database, with the use of mpeg 7 we can overcome this issue.


1. Eidenberger, H. (2004), “A Video Browsing Application Based on Visual MPEG-7 Descriptors and Self-Organising Maps”, [Online]. 6(3), Available at: www.ims.tuwien.ac.at/media/documents/publications/jfs2004.pdf [Accessed on 22 April 2010].

2. Dollar, M. (2007) ”The MPEG-7 Multimedia Database System (MPEG-7 MMDB)” [Online] Available at: http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V0N-4R5DYW0-1&_user=10&_coverDate=09/30/2008&_rdoc=1&_fmt=high&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=1308196117&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=9fdd1c0220d6316dc26bd663ee041953 [Accessed on 28 March 2010].


CONTENT-BASED MANAGEMENT AND RETRIEVAL OF MUSIC”, [Online]. Available at: http://ismir2004.ismir.net/proceedings/p010-page-48-paper227.pdf [Accessed on 20 April 2010].

4. Doller,M. (2006), MPEG-7 Meets Multimedia Database Systems, [Online]. 1(1). Available at: http://docs.google.com/viewer?a=v&q=cache:vX8qAJHsTLEJ:citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.

[Accessed on 15 April 2010]

5. Mallorca, D. P. (2004), MPEG 7 OVERVIEW, [Online] Available at: http://mpeg.chiariglione.org/standards/mpeg-7/mpeg-7.htm [Accessed on 17 March 2010].