The relational database model is based on relational algebra, which is related to first-order logic and the algebra of sets. The relational model is powerful enough to describe most organizational and administrative tasks of modern society.
In our century, the nature of documents and information is changing; more information is represented by images, films and unstructured text.
When examining multimedia databases, we have to consider subsymbolical AI, algorithms from signal processing, image recognition, high-dimensional indexing and machine learning. This is what World Scientific's latest book on "Intelligent Big Multimedia Databases" seeks to do.
The history of multimedia databases began with the use of photography to record known criminals as early as the 1840s.
Early applications of multimedia database management systems only employed multimedia for presentational requirements: a sales order processing system may include an online catalog that includes a picture of the offered product. However, this simple extension of the relational model is insufficient when handling multimedia information.
When handling multimedia information, we have to consider digital data representations and explore questions regarding how these data can be stored and manipulated:
- How to pose a query?
- How to search?
- How can information be retrieved?
A multimedia database provides more functions than are available in the traditional form of data representation. One example of such a function is content-based image retrieval (CBIR). An image or drawn user input serves as a query example; as a result, all similar images should be retrieved. The size of a multimedia object may be immense. For the efficient storage and retrieval of large amounts of data, a clever method for encoding information using fewer bits than the original representation is essential. Feature extraction is a crucial step in content-based image retrieval. The set of features represents relevant information about the input data in a certain context.
Feature extraction is related to compression algorithms and is frequently based on the transform function such as image pyramid or discrete fast wavelet transform for images.
We describe information retrieval techniques and essential statistical supervised machine learning algorithms. The book is based on the idea of hierarchical organization of information processing and representation, such as the wavelet transformation, the scale space, the subspace tree and deep learning.
The best-known example is the content-based image retrieval (CBIR) on the basis of automatically derived features. An image or some drawn user input serves as a query example, and all similar images should be retrieved as results. An image query is performed through the generation of a weighted combination of features, and through its direct comparison with the features stored in the database.
A similarity metric (e.g. the Euclidean distance) is then used to find the nearest neighbours of the query example in the feature vector space. Feature extraction is the crucial step in content-based image retrieval. The extracted features of a CBIR system are mapped into points in a high-dimensional feature space, and the search is based on points that are close to a given query point in this space. For efficiency, these feature vectors are precomputed and stored.
The metric tree index tree operates efficiently only when the number of dimensions is small (< 10). The growth in the number of dimensions has negative implications in the performance of multidimensional index trees; these negative effects are also known as the "curse of dimensionality."
Traditional indexing of multimedia data leads to dilemma. Either the number of features has to be reduced or the quality of the results in unsatisfactory, or approximate queries is preformed leading to a relative error during retrieval.
While there are relatively efficient approximate similarity search algorithms, it is widely supposed that the exact search suffers from dimensionality. Thus, solving the problem in the most general case for an arbitrary dataset seems impossible. However the problem of rapid exact search of large high-dimensional collections of objects is an important problem with applications in many different areas (multimedia, medicine, chemistry, biology, etc).
We show how the recursive application of the lower bounding postulate can be used to overcome the curse of dimensionality for certain cases of points equally distributed by subspace trees. A high-dimensional space is divided into low-dimensional sub-spaces. In the low- dimensional sub-spaces, lower bounding postulate can be successfully applied.
The most significant aspect of the book is the idea of the hierarchical organization of information processing and representation, such as the wavelet transformation, the scale space, the subspace tree and deep learning. The resulting idea is of how these aspects can overcome the curse of dimensionality to develop big multimedia databases.
This research was supported in parts by the grant Ciencia e Tecnologia (Portugal) through the programme PTDC/EIA-CCO/119722/2010.
This set is sold in major book stores at US$115 / £76. More information on the book can be found at http://www.
To schedule an author interview, please contact Jason CJ Lim at firstname.lastname@example.org
About World Scientific Publishing
World Scientific Publishing is a leading independent publisher of books and journals for the scholarly, research and professional communities. The company publishes about 500 books annually and more than 120 journals in various fields. World Scientific collaborates with prestigious organisations like the Nobel Foundation, US National Academies Press, as well as its subsidiary, the Imperial College Press, amongst others, to bring high quality academic and professional content to researchers and academics worldwide. To find out more about World Scientific, please visit http://www.
For more information on the book, contact Jason CJ at email@example.com