Multimedia Database Management Systems (The Springer ...

Good...:
This book is good and it contains almost all the concepts of this area published till the date of the publication of this book. I would mainly like to comment Mahabala, the first reviewer of this book. He was my Ph.D. guide in IIT Madras. I spent eight years with him and at the end he simply sent me out without allowing me to submit the thesis (Another Professor Khemani helped him in this). Please please do not believe Prof. Mahabala's words. He does not know any thing in modern computer science but he speakes as if he knows every thing. May be he is feeling that he is eligible for receiving Turing award!!! He always worried how to get admission for his daughter's (who got MCA degree in India) MS admission in the University of Maryland. Of course finally he managed. I hate this Mahabla and his comments. The best filed for him is Indian politics and I am sure he will shine as a politician in India!!!

All that an IT professional has to know about Multimedia:
Multimedia (MM) combined with Internet has revolutionized the concept of information particularly in data communication. Before many in the developing countries could fully understand the power of the Internet, and get ready to use text oriented data, they are told that the current Internet is not suited for multimedia! One is asked to evaluate a PC one intends to buy, from the multimedia angle. Multimedia is no longer a research topic, but a fact of life in the IT dominated world of today. The book by Dr B Prabhakaran, who did his Ph. D. in IIT, Madras (now Chennai) will satisfy the needs of a reader who wants to understand the basics of MM as well as the state-of-the-art as of today. Even though the book is written by a researcher, it presents the material in a manner, easily understood by an IT professional. Ability to code information in any form: text, image, graphics, voice, and movies, as binary files is the basis for MM. One will however, have to use appropriate encoders and presentation devices. Problems in handling MM are its large size, and its real-time nature for presentation purposes. Memory requirement for MM is however reduced by using certain compression (associated decompression) techniques. MM databases are classified as Orchestrated (composition of stored objects) or Live (generated by devices such as microphone or Video camera). They can also be classified as Discrete (time independent such as text, graphics or image), or Continuous such as video, and audio. MM storage and management involves huge files (of the order of Gigabytes), supporting simultaneous synchronized access to multiple files (audio and video), and supporting large disk access bandwidth. Book discusses various ways of allocating space on the disk to achieve maximum bandwidth. Once a MM presentation at the client's end has started, there can be no interruption due to network delay or noticeable gaps due to lost packets. A MMDB server has to have the freedom to admit a new request based on its ability to deliver to the network MM data meeting the real-time requirement. The network in turn should guarantee delivery of the data with requisite Quality of Service (QoS) as specified by the maximum delay (with a specified jitter) and probability (low enough) of the loss of packets. Since we want to retrieve MM from a server across the network based on our information need, it is interesting to think in terms of how we will specify our need. Obviously, we want support for following type of queries: Get me the movie in which this scene appears, or this song is played. Show me the movie 30 minutes after its start Show me the image where President Yeltsin is seen to the left of President Clinton. Show me the video where the river changes its course. We cannot as yet handle all these query types. In order to handle some of these queries, one has to associate with each MM data Metadata, which is data about data. Metadata such as author's name or date of recording has to be entered with the MM data. Certain other types of metadata which are based on the content can be obtained by processing the MM data. Indexing an MM database has to be based on the metadata. Content-based retrieval is in its infancy and one can expect involvement of many disciplines such as: statistics, image and speech recognition, automated reasoning, data mining, etc. In a way, content-based retrieval is the ultimate research problem. The book covers the essence of all topics relevant to making MM a commodity service, without burdening the user with unnecessary details. Reviewer would like to urge all IT professionals to get comfortable with multimedia by reading this highly readable book. The book can also be used as a text book for a course on MM.