Multimedia Data Mining Trends and Challenges

Multimedia Data Mining Trends and Challenges

Janusz Swierzowicz (Rzeszow University of Technology, Poland)
DOI: 10.4018/978-1-60566-014-1.ch131
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

The development of information technology is particularly noticeable in the methods and techniques of data acquisition. Data can be stored in many forms of digital media, for example, still images taken by a digital camera, MP3 songs, or MPEG videos from desktops, cell phones, or video cameras. Data volumes are growing at different speeds with the fastest Internet and multimedia resource growth. In these fast growing volumes of digital data environments, restrictions are connected with a human’s low data complexity and dimensionality analysis. The article begins with a short introduction to data mining, considering different kinds of data, both structured as well as semistructured and unstructured. It emphasizes the special role of multimedia data mining. Then, it presents a short overview of data mining goals, methods, and techniques used in multimedia data mining. This section focuses on a brief discussion on supervised and unsupervised classification, uncovering interesting rules, decision trees, artificial neural networks, and rough-neural computing. The next section presents advantages offered by multimedia data mining and examples of practical and successful applications. It also contains a list of application domains. The following section describes multimedia data mining critical issues, summarizes main multimedia data mining advantages and disadvantages, and considers some predictive trends.
Chapter Preview
Top

Multimedia Data Mining State Of The Art

One of the results of the inexorable growth of multimedia data volumes and complexity is a data overload problem. It is impossible to solve the data overload issue in a human manner; it takes strong effort to use intelligent and automatic software tools for turning rough data into valuable information and information into knowledge.

Data mining is one of the central activities associated with understanding, navigating, and exploiting the world of digital data. It is an intelligent and automatic process of identifying and discovering useful structures in data such as patterns, models, and relations. We can consider data mining as a part of the overall knowledge discovery in data processes. Kantardzic (2003, p. 5) defines data mining as “a process of discovering various models, summaries, and derived values from a given collection of data.” It should be an iterative and carefully planned process of using properly analytic techniques to extract hidden, valuable information.

Key Terms in this Chapter

MP3: MPEG audio coding standard layer 3. The main tool for Internet audio delivery.

X3D: Open Standards XML (Extensible Markup Language) enabling 3D (dimensional) file format, real-time communication of 3D data across all applications, and network applications.y

Multimedia Data Mining: Extracting interesting knowledge out of correlated data contained in audio, video, speech, and images.

MPEG: Motion Picture Engineering Group.

MX: An XML (Extensible Markup Language)-based formalism for the representation of the music pieces.

MPEG-4: Provides the standardized technological elements enabling the integration of the production, distribution, and content access paradigms of digital television, interactive graphics applications, and interactive multimedia.

Data Mining: An intelligent and automatic process of identifying and discovering useful structures such as patterns, models, and relations in data.

Image Mining: Extracting image patterns, not explicitly stored in images, from a large collection of images.

MPEG-7: Multimedia Content Description Interface.

Complete Chapter List

Search this Book:
Reset