Oge Marques and Borko Furht
Department of Computer Science and Engineering
Florida Atlantic University
Boca Raton, FL, USA
The field of distributed multimedia systems has experienced an extraordinary growth during the last decade. Among the many visible aspects of the increasing interest in this area is the creation of huge digital libraries accessible to users worldwide. These large and complex multimedia databases must store all types of multimedia data, e.g., text, images, animations, graphs, drawings, audio, and video clips. Video information plays a central role in such systems, and consequently, the design and implementation of video database systems has become a major topic of interest.
The amount of video information stored in archives worldwide is huge. Conservative estimates state that there are more than 6 million hours of video already stored and this number grows at a rate of about 10 percent a year . Projections estimate that by the end of 2010, 50 percent of the total digital data stored worldwide will be video and rich media . Significant efforts have been spent in recent years to make the process of video archiving and retrieval faster, safer, more reliable and accessible to users anywhere in the world. Progress in video digitization and compression, together with advances in storage media, have made the task of storing and retrieving raw video data much easier. Evolution of computer networks and the growth and popularity of the Internet have made it possible to access these data from remote locations.
However, raw video data by itself has limited usefulness, since it takes far too long to search for the desired piece of information within a videotape repository or a digital video archive. Attempts to improve the efficiency of the search process by adding extra data (henceforth called metadata) to the video contents do little more than transferring the burden of performing inefficient, tedious, and time-consuming tasks to the cataloguing stage. The challenging goal is to devise better ways to automatically store, catalog, and retrieve video information with greater understanding of its contents. Researchers from various disciplines have acknowledged such challenge and provided a vast number of algorithms, systems, and papers on this topic during recent years. In addition to these localized efforts, standardization groups have been working on new standards, such as MPEG-7, which provide a framework for multimedia content description.
The combination of the growing number of applications for video-intensive products and solutions - from personal video recorders to multimedia collaborative systems - with the many technical challenges behind the design of contemporary video database systems yet to be overcome makes the topics discussed in this Handbook of extreme interest to researchers and practitioners in the fields of image and video processing, computer vision, multimedia systems, database systems, information retrieval, data mining, machine learning, and visualization, to name just a few (Figure 1.1).
Figure 1.1: Visual Information Retrieval blends together many research disciplines.
In this chapter we present a general overview of the central topic in this Handbook: video databases. Its main goal is to introduce basic concepts behind general-purpose database systems and their extension to multimedia database systems, particularly image and video database systems. Section 2 introduces basic database concepts and terminology. In Section 3 we briefly outline the main steps of the database design process. Section 4 extends the discussion from general databases to multimedia databases and their own particular requirements and characteristics. Section 5 narrows down the field even further, focusing on the aspects that are specific to image and video databases. The main goals of this Handbook are outlined in Section 6. Finally, Section 7 provides the reader with an overview of the other chapters of this Handbook and how they have been organized.