Multimedia Databases: An Object-Relational Approach / Edition 1

The topic, multimedia databases, itself is interesting and poses a number of challenges for the database developer. The reader faces challenges that arise from integrating knowledge from a number of different fields such as multimedia, networks and communication systems, HCI and human sensory systems as well as database concepts. One of the motivations for writing this book came from the experience of needing to refer to numerous textbooks from different disciplines when developing multimedia databases.

Who Should Read this Book?

The primary audience for this book is those whose working life will be increasingly influenced by a need to know more about multimedia databases. This would include those working in web technology and e-commerce where many of the issues are now recognized to bedatabase problems. Others may be involved in the growth of applications in this field, for example, media on demand, surveillance systems, GIS and medicine. Furthermore, it is likely that in the future most traditional commercial database systems will incorporate multimedia data types. For example, personnel systems will include images of employees, SOP systems will have images of products and videos of usage. Visualizations are also being incorporated into decision support systems derived from databases. Multimedia data include text so that the management of large document collections are also within the scope of this book.

This book could also be of interest and benefit to students of computing and information systems in the later stages of their degree programs. Since interest in multimedia is growing there are also students, now studying degree programs for multimedia information systems, who would find the text useful. This audience may have a good knowledge of relational databases but need an introduction to additional features such as large binary objects, methods and user-defined types available in SQL3. The object/relational approach can exploit features of PL/SQL (procedural SQL) to provide methods for the new multimedia object types.

The book can be followed within a second- or third-level course in advanced database technologies. It also includes examples and exercises to illustrate the principles and promote understanding. SQL code can be used in practical sessions and is supplied on the accompanying CDROM and the website. The teaching approach is a blend of theory and practice. Wherever possible readers should be encouraged to try the examples within an appropriate database that supports multimedia data types. There are also examples of websites to illustrate techniques and applications. Two case studies are included in the text and there are further examples on the CDROM. SQL code is provided to implement simple multimedia databases for image, text and video data.

The production of multimedia systems themselves is not covered. For example, only the database aspects of development of application areas such as multimedia learning systems is mentioned. Details of the generation of the media by scanning or photographic systems are not included but the significance of different compression standards is discussed.

Multimedia databases are the result of the conjunction of developments in a number of areas. Advances in data capture and compression have allowed vast collections of multimedia data to be stored, while SQL3 provides a language that deals with large binary objects that can be used to store it. Object-oriented and object relational database design gives an insight into the effective realization of databases for this kind of data but up to now the expertise in this area has tended to rest in isolated research enclaves rather than to be generally available to database practitioners.

There had been a rapid development of the technology in databases and multimedia. In terms of databases the introduction of SQL3 in 1999 has corresponded with a number of vendors implementing features to store and manipulate large binary objects within object-relational databases. The development of technologies that provide content-based retrieval of media objects is highly significant. Early multimedia databases tended to be designed using object-oriented methods that were often application dependent and did not provide generic capabilities especially in terms of query languages.

How to Read this Book

The book overall is intended to be read in sequence but since readers may come from a background of either multimedia or databases they may skip appropriate sections in the earlier chapters. For example, readers with a background in multimedia will be familiar with much of the material in Chapter 2, while those from a database background will be familiar with much of Chapter 1 and the early sections of Chapter 4. There are separate chapters (10, 11, 12) dealing with database systems about text, image and real-time media (audio and video). These contain specialized information which the reader may chose to kread over lightlym if it is an area that is not relevant to their interests. For example, Chapter 11 on image databases contains information derived from the disciplines of image processing and computer vision that is quite technical.

In the main text where examples of SQL statements are included SQL keywords are shown in upper case, although Oracle ignores upper and lower case except when testing for conditions involving character strings (such as department # kAccountsm). Table, column and object names are generally in lower case, except where the column name is made up of several words such as numberOfFrames.

What is Included in the CD-ROM?

There is an accompanying CD-ROM on which is stored a series of exercises and solutions that relate to the various chapters and exercises of the book. The solutions to the exercises included consist of SQL and PL/SQL code for Oracle, MySQL and DB2 implementations of examples of multimedia databases. In addition, there are a number of case studies that provide additional practice in the implementation of MMDBMS.

The purpose of the CD-ROM is to provide examples that you can use to interact directly with the database management system. The CD-ROM is organized in chapters which contain the Oracle SQL*PLUS code referred to in the book text. There are separate exercise files for DB2 and MySQL.

The exercises usually involve the preparation and execution of SQL statements. The files contain a copy of most of the statements from the text in the book. You can simply type in the statements into an SQL editor or select and paste the SQL text. In the chapters some of the SQL statements are named. For example in Chapter 4

CREATE TABLE department

(department_number CHAR(4) CONSTRAINT prim_dept PRIMARY KEY,

department_name VARCHAR2(10 )

prim_dept

The same name prim_dept is used in the corresponding exercise files so that you can locate the name of any particular query using a find without having to scroll through the whole text.

What Software is Used in this Book?

Most of the SQL examples used in the book are based on Oracle, versions 8i and 9i. The Oracle Corporation appears currently to be producing new versions and upgrades of their products at a rapid rate. However, for the majority of this book you do not need to worry about which version of Oracle you are using. Where these differ in terms of multimedia data this is made clear in the text. In addition, there are examples for DB2, mySQL and XML. In the case of DB2 this is a database management system with significant advantages and capabilities in the retrieval of multimedia data. However, since this text is also focused on SQL3 the procedural code examples are mainly written in Oraclems PL/SQL as this is closer to the SQL standard and development with DB2 tends to involve using a host language.

Acknowledgments

This book would not have been possible without the contribution of many people who were involved directly or indirectly in the production of the book.

First, I would like to thank my friend Osei Adjei for very thoroughly reviewing large sections of the manuscript, especially Chapters 5 and 12. Second, I benefited enormously from comments and discussions over several years with my former colleague Iain McClaren and the members of the Open University database team, Mike Newton, Hugh Robinson, Kevin Waugh, Steven Self and, of course, Hugh Darwen.

I would also like to thank Judith Sewell for patiently reviewing early drafts and raising many issues even though she would not consider herself a ldatabase personn.

I obtained considerable technical support with the various Oracle installations I used for the development of the practical exercises from my colleagues John Mullins and Frank Hines at the IT Center, Thames Valley University.

I should also like to thank and acknowledge Alex Murray, the Web Master at Berry Brothers & Rudd for providing information and support in the development of the Fine Wine Shop case study and for providing such an interesting website.

The ideas for the protozoa database case study came from Dr David Roberts of the Natural History Museum and I would like to thank and acknowledge his help and support with the case study and for introducing me to the fascinating world of protozoa.

Finally, I am grateful to everyone at Addison-Wesley, especially Viki Williams my editor, and Tessa Fincham for all their encouragement and work.