International Journal of Computer Science Trends and Technology (IJCST) Volume 3 Issue 1, Jan-Feb 2015

RESEARCH ARTICLE

OPEN ACCESS

Metadata Based Approach For Web Multimedia Model

Siddu P. Algur1, Basavaraj A. Goudannavar2, Prashant Bhat3
Professor1, Research Scholar2*, Research Scholar3
Department of Computer Science,
Rani Channamma University, Belagavi
Karnataka - India

ABSTRACT
Due to the rapid growth of digital information over the Internet, it is becoming very important to extract useful information
from web multimedia data. The WWW (World Wide Web) provides a simple and effective means for users to retrieve
multimedia data on the Internet. Nowadays web multimedia data are used in a much more natural and mature way. Rapid
increase in the huge amount of multimedia data over the internet, there is emerging trend to study, modeling and retrieve, and
mining the multimedia data from the internet. Due to complex and unstructured nature of the web multimedia data, it is difficult
for mining and processing effectively. To overcome from this problem, this paper proposes a new web multimedia data model
for presenting web multimedia data components integrating large amounts of data of different types such as text, images, video
and audio, with its metadata values for mining and processing web multimedia data effectively.
Keywords:- Web Multimedia, Web Multimedia Model, Multimedia Metadata.

I. INTRODUCTION
The advances in computer and network infrastructure
together with the fast evolution of digital data are resulted in
the growth of attention to the digital multimedia development
over the web. The scientific community has increased the
amount of research into new technologies, with a view to
improving the digital web multimedia utilization: its archiving,
indexing, accessibility, acquisition, store and even its process
and usability. All these parts of the web multimedia utilization
lead to the necessity of the extraction of all important
information of web multimedia by means of lack of metadata.
Over the last few years there has been observed an
increasingly use of web multimedia data in the collection of
web digital libraries. This fact automatically requires not only
the transmission but also the store and the process of the
relevant information are essential [1].
Web digital libraries evolved from traditional libraries
during the information age, have become important platforms
for information retrieval from the Internet. However, due to
the rapid expansion of multimedia data in the collection of
web digital libraries, such as text documents, video, audio and
image data, traditional information retrieval approaches
became inadequate to the management of large unstructured
multimedia data. The reasons are-firstly, support of interaction
with multimedia data is insufficient. Users are usually passive

ISSN: 2347-8578

roles only provided with one-way access to multimedia data.

Even though web multimedia data carry a rich amount of
information, difficulty has been acknowledged in systematic
management of such unstructured data. Secondly, there are
models that associate web multimedia data with the keywords
for information management and retrieval, where web
multimedia data are often stored as object attributes with low
granularity, which is prone to cause data redundancy and can
hardly support content-level information retrieval. Thirdly,
distributed structure, considered as one of the optimal
solutions for developing search engines [2][3], with
advantages in information retrieval from massive data, is also
facing a big challenge in the efficient organization and
management of web multimedia data. For web digital libraries,
an effective way of managing multimedia data has to be able
to automatically handle the embedded information in its
content. In traditional libraries, users can find desired objects
from the multimedia archives by using indexing tools like
cards. By analogy with that, a structured index can also be
established for the content of web multimedia data in web
digital libraries, so that the embedded information can be
exposed for retrieval, and un-ordered multimedia data can be
translated into index-based data which is more suitable for
storage and access control with fine granularity. Therefore,
proper data models are needed for organizing and managing
massive web multimedia data, which should be able to
provide web multimedia information retrieval service for web

www.ijcstjournal.org

Page 11

International Journal of Computer Science Trends and Technology (IJCST) Volume 3 Issue 1, Jan-Feb 2015
digital libraries by means of multimedia metadata and multilevel semantic data mining. In response to such need, this
paper proposes an efficient data model for web multimedia
data by means of multimedia specific metadata.

II. RELATED RESEARCHES

A framework for customizable metadata authoring that
addresses the issues and challenges is described by the authors
Valentina Malaxa and Ian Douglas [3], and observed that,
there is an obvious requirement for metadata as a means for
exchange, discovery, and sharing of digital objects. Metadata
is used by the applications, repositories, content creators, and
system developers. In environments in which there are
potentially huge amount of objects available, the metadata
must be sufficient, effective, accurate, and complete. This
observation has identified challenges related to users needs,
metadata standards, and learning environment and proposed
solution these challenges. Users require better interfaces to
support their goals and better guidance than that available in
standards documentation. During this research, collaborative
editing was identified as being one of the critical features in
achieving quality metadata.
Multimedia storage structure and data organization
methods are the key technology affecting the development of
the digital library [4]. This designs a novel data model, and
efficiently realizing the organization of multimedia data. The
innovation point expresses as: It provides different fine
granularity control on the media content, and sharply reduces
the information redundancy within the media and among the
medias, meanwhile better support the content-based retrieval
with enhances the ability of abstraction and generalization of
the model, which can support a variety of media objects,
studio, or the performance of the query.
Andreas Pleub [5] presented a model-driven
development approach for interactive multimedia applications.
It consists of the Multimedia Modeling Language (MML) and
automatic transformations into code skeletons for different
platforms. MML is platform independent and bases on
existing modeling approaches like UML, user interface
modeling approaches, and multimedia modeling. The resulting
modeling language provides support for integration of
software design, media design, and user interface design. In
addition, it allows modeling advanced concepts of media
objects, such as interactivity and dynamic alterations. Both,
models and transformations, are defined in compliance to the
concepts of Model-Driven Engineering.
The author Hasan Harasis [6], proposed an efficient
interactive model for multimedia application. This model
combines the advantages of aspect programming and objects
oriented programming. As a result, our proposed model

ISSN: 2347-8578

applies a structure for multimedia applications that has good

reutilisation of code in other applications and objects

storage in databases. In this approach, the time required for
realization of an entire multimedia application was decreased
and the size of the executing file became smaller. Using this
model it is possible to develop very efficient multimedia
applications [6].
The authors Patrizia Asirelli [7], proposed an initial
prototype version of the infrastructure has been developed that
integrates the prototype versions of the MPEG-7 feature
extraction and database units. This prototype was
demonstrated some of the technical challenges faced in
integrating multimedia metadata. On the whole, the
architecture proposed here enables media to be combined and
managed [7]. In addition valuable semantic services can be
supported, such as semantic search and retrieval, algorithm
discovery and application and semantic annotation.
The emerging standards of SMIL, SVG and MPEG-4
provide a new process for authoring and presenting
multimedia documents, also known as multimedia integration
and synchronization. It introduces a new multimedia type
which enables the integration of a rich set of media into more
complex structures and provides news interaction capacity in
multimedia presentations. Macromedia Director, MPEG-4 and
X-SMIL can be used for this purpose. These tools provide a
sophisticated high level graphical editing interface like
timeline and layout views for integrating and synchronizing a
set of media. However all of them still require the author a
long and relatively complex authoring process, especially
when fine-grained synchronization is desired. For example, an
author wants to display a text introducing a character in a
video when this character occurs on screen. The authoring
process requires the manual determination of the temporal
information, the begin time and the stop time, of the
appearance of the character in the video and the absolute
temporal placement of the text along with this temporal
information. The difficulty involves the effort taken to
determine temporal information inside the video, because
current multimedia authoring tools do not support media
content analysis and visualization of high level content
structures of video. It becomes more complex when making a
hyperlink on the video character or making a text following
the video character because the author needs to determine not
only temporal information but also spatio-temporal
information of the character. At this point, the most important
standard multimedia integration model SMIL fails to integrate
these specification needs. However, it has been observed that,
following are the several limitations of the existing models [8].
The existing models do not address how to extract
the search items from the web multimedia files/data.

www.ijcstjournal.org

Page 12

International Journal of Computer Science Trends and Technology (IJCST) Volume 3 Issue 1, Jan-Feb 2015

Extraction of audio and visual information from web

multimedia files is time consuming.
Retrieval of video clip segment from web multimedia

Web
Multimedia
Data

is very tedious work.

Retrieval of images from web multimedia data is
requires more time.
Manual extraction of the features of web multimedia is time
consuming

Decomposition
Process

III. THE PROPOSED WEB MULTIMEDIA

DATA MODEL
The objective of the proposed model is to extend the
limitations to the existing multimedia model. To meet scope
of the proposed model, the model extend its limit to
multimedia object-specific metadata. The web multimedia
data is composed of four basic objects- text, image, audio and
video. Each web multimedia object has metadata values, a
logical structure and an interpretation of the content for that
value. The metadata value of a web multimedia object must be
described in the web multimedia data model. The Fig.1 shows
the proposed web multimedia model, which describes
multimedia objects and its structure with metadata repository.
This model can be used to decompose the web multimedia
data into its basic objects- text, image, audio and video, also
the model extracts web multimedia object specific metadata
and stores in a repository using appropriate available tools, so
that the mining task will be effective and efficient for
knowledge discovery. There could be sufficient metadata or
insufficient metadata for multimedia objects. If sufficient
metadata are available for the web multimedia objects, then
knowledge discovery will be more effective. There are many
tools available to decompose the web multimedia object
meaning that, it is possible to retrieve individual
components/object from multimedia data. To separate and
extract audio and video from audio-visual data, a tool called,
Oxelon Media Converter, can be used, also to extract images
from video data; XVideo Converter tool is available. The
web multimedia text object has text specific metadata such as,
language of the text, font information, quality of the text,
number of lines and characters etc. The web multimedia
image object has image specific metadata such as, image
format information, resolution, dimension, compression, color
mode etc. To extract text metadata, a Metadata Extraction
Tool is available which is developed by National Library of
New Zealand.

ISSN: 2347-8578

Text

Audio

Video

Image

has

has

has

has

Text
specific
Metadata

Audio
specific
Metadata

Video
specific
Metadata

Image
specific
Metadata

Metadata Extraction process

Multimedia
Metadata
Repository

Multimedia Processing
Metadata

Fig.1: Proposed Web Multimedia Model

The web multimedia video object has three types of video
specific metadata- technical metadata, web metadata as well
as descriptive metadata. The technical metadata includes
format and structural information about the video object, such
as- format, bit rate, resolution, aspect ratio etc. The web
metadata includes both implicit and explicit metadata such asURL and website information, ratings, likes, dislikes
comments etc. The descriptive metadata of web multimedia
object includes descriptions and summary about the content of
the object. To extract video specific metadata, MediaInfo,
InfoExtractor tools are available. Similarly, the web
multimedia audio object has audio specific metadata such asaudio track information, audio bit rate, channel information,
duration of the audio etc. To extract audio specific metadata,
MediaInfo tool is helpful. Further, all types of metadata are

www.ijcstjournal.org

Page 13

International Journal of Computer Science Trends and Technology (IJCST) Volume 3 Issue 1, Jan-Feb 2015
stored in a repository for processing and mining web
multimedia data.

IV. A DOMAIN SPECIFIC APPLICATION OF

PROPOSED MODEL
The proposed web multimedia metadata model has been
designed as the assembly of modules dedicated to the
description of separable aspects of an object's description.
Some modules are centered on web multimedia object specific
metadata. In the following, we concentrate on metadata for
news video object; however, these characteristics can be
generalized to other video domains. The news domain
application of proposed model is presented in Fig.2. The
domain specific model use to present the main concepts with
their attributes and relationships.
The news domain describes the common features of
multimedia objects. In this section we focus on news video
object, by showing how objects of video data have relations
among remaining multimedia objects (one-to

many relationship) using descriptive metadata. For example,

the audio in video object may have relationship with
multimedia audio object; the texts in video object may have
relationship with multimedia text object; the resolution of
video object has relationship with multimedia image object,
etc. For the personalization system to purpose, we require
acquisition of structural and satisfied metadata describing the
news sectors. In our model, this is captured by reading the
close description text metadata. A more suitable approach of
obtaining this metadata is directly from the news providers.
User Metadata Since we use cognitive filtering, our user
profile corresponds closely to metadata available for each
news segment (i.e., user metadata are nearly identical to
content metadata). We thereby choose a descriptive user
profile constructed by observing user manipulations of
previously constructed news story sequences.

Web Multimedia Data

Sports Video Clip

Entertainment Video Clip

Video Category N

News Video Clip

Text

Audio

Anchoring
Language
Font
Quality
Line Terminals
No. of Lines
No. of Character

Image

Video

Discussion

Track
Duration
Channel
Bit

rate

- No. of Persons
- Theme and concepts
of news
- Date and Time
- Subtitles Yes/No
(Text)

- No. of Persons
- Theme of the
Discussion
- Date and Time
Information
- Subtitles Yes/No
(Text)

Format
Resolution
Dimension
Color mode
Compression

Multimedia
Metadata
Repository

Multimedia Processing using

Metadata
ISSN: 2347-8578

www.ijcstjournal.org

Page 14

International Journal of Computer Science Trends and Technology (IJCST) Volume 3 Issue 1, Jan-Feb 2015

Fig.2: Metadata based Web Multimedia Model for News Domain

For these users descriptive metadata such as audio, text,
and video also have the additional attributes of playout times
and required bandwidth associated with them. It thus becomes
necessary for the system to be aware of the timing and
bandwidth characterizations of a multimedia object when

V. CONCLUSIONS
In this paper, we proposed an effective and efficient web
multimedia data model by reviewing the existing multimedia
models and their limitations. Using different available tools,
the proposed model decomposes the multimedia data into its

requested, so that appropriate resources can be reserved to

ensure the timely delivery of data. The use of metadata to
support database browsing or personalization functions is
significant due to the potential bandwidth used in video
delivery.
basic components- text, image, audio and video and extracts
respective appropriate metadata for basic components for
multimedia processing and mining purpose. To demonstrate
the proposed model, a domain specific illustration is presented
for news video domain.

REFERENCES
[1] Dim P. Papadopoulos, , Vicky S. Kalogeiton, Savvas A. Chatzichristofis, Nikos Papamarkos, Automatic summarization
and annotation of videos with lack of metadata information Elsevier http://www.elsevier.com Volume 40, Issue 14, 15
October 2013, Pages 57655778.
[2] Brunelli, R., Mich, O., Modena, C.M.: A survey on the automatic indexing of video data. Journal of Visual
Communication and Image Representation. 10(2), 78-112(1999).
[3] Valentina Malaxa, Ian Douglas A Framework for Metadata Creation Tools, Interdisciplinary Journal of Knowledge and
Learning Objects Volume 1, 2005.
[4] Lin Liu, A Data Model with Multimedia Retrieval, The Third International Conferences on Advances in Multimedia,
IARIA, 2011. ISBN: 978-1-61208-129-8. Advanced Science and Technology Letters Vol.48 (CIA 2014), pp.47-50. ISSN:
2287-1233.
[5] Andreas Pleu , Model-Driven Development of Interactive Multimedia Applications Towards Better Integration of
Software Engineering and Creative Design, Mnchen, den 25. Februar 2009.
[6] Hasan Harasis, New Interactive Multimedia Application Model Using Objects Storage In A Database, Journal of
Theoretical and Applied Information Technology, 20th February 2014. Vol. 60 No.2, ISSN: 1992-8645, E-ISSN: 18173195.
[7] Patrizia Asirelli, Suzanne Little, Massimo Martinelli and Ovidio Salvetti, MultiMedia Metadata Management: a Proposal
for an Infrastructure, MultiMedia Metadata Management: a Proposal for an Infrastructure, submitted SWAP 2006.
[8] Tien Tran-Thuong, Ccile Roisin, Multimedia Modeling Using MPEG-7 for Authoring Multimedia
Integration, MIR03, November 7, 2003, Berkeley, California,
USA. 2003 ACM 1-58113-778-8/03/0001.

BIOGRAPHIES
Mr. Basavaraj A. Goudannavar is pursuing Ph.D programme
in
Computer
Science
at
Rani
Channamma
University
Belagavi,
Karnataka, India. He received BCA and
MCA degrees from Karnatak University,
Dharwad, Karnataka, India, in 2005 and
2008 respectively. His research interest
includes Data Mining, Web Mining, Web
multimedia mining, and Knowledge
discovery techniques. He published 02
research papers in
International Journals and International conferences He has
attended and participated in International and National
Conferences and Workshops in his research field.

ISSN: 2347-8578

Mr. Prashant Bhat is pursuing Ph.D programme in

Computer
Science
at
Rani
Channamma University Belagavi,
Karnataka, India. He received B.Sc
and M.Sc (Computer Science)
degrees from Karnatak University,
Dharwad, Karnataka, India, in 2010
and 2012 respectively. His research
interest includes Data Mining, Web
Mining,
web multimedia mining and Information Retrieval from the
web and Knowledge discovery techniques, and published
research papers in International Journals. Also he has
attended and participated in International and National
Conferences and Workshops in his research field.

www.ijcstjournal.org

Page 15

International Journal of Computer Science Trends and Technology (IJCST) Volume 3 Issue 1, Jan-Feb 2015
Dr. Siddu P. Algur is working as Professor, Dept. of Computer
Science, Rani Channamma University,
Belagavi, Karnataka, India. He received
B.E. degree in Electrical and Electronics
from Mysore University, Karnataka, India,
in 1986. He received his M.E. degree in
from NIT, Allahabad, India, in 1991.He
obtained Ph.D. degree from the Department
of P.G. Studies and Research in Computer
Science at Gulbarga University, Gulbarga.
He worked as Lecturer at KLE Societys College of Engineering
and Technology and worked as Assistant Professor in the
Department of Computer Science and Engineering at SDM
College of Engineering and Technology, Dharwad. He was
Professor, Dept. of Information Science and Engineering,
BVBCET, Hubli, before holding the present position. He was
also Director, School of Mathematics and Computing Sciences,
RCU, Belagavi. His research interest includes Data Mining,
Web Mining, Big Data and Information Retrieval from the web
and Knowledge discovery techniques. He published more than
40 research papers in peer reviewed International Journals and
chaired the sessions in many International conferences.

ISSN: 2347-8578

www.ijcstjournal.org

Page 16