Database Trends and Directions: Current Challenges and Opportunities
George Feuerlicht1,2
1 Department of Information Technology,
University of Economics, Prague, Czech Republic
2 Faculty of Engineering and Information Technology,
University of Technology, Sydney
Abstract. Database management has undergone more than 4 decades of evolution producing vast range of research and extensive array of technology solutions.
The database research community and software industry has responded to numerous challenges resulting from changes in user requirements and opportunities presented
by hardware advances. The relational database approach as represented by SQL databases has been particularly successful and one of the most durable paradigms in
computing. Most recent database challenges include internet-scale databases - databases that manage hundreds of millions of users and cloud databases that use novel techniques for managing massive amounts of data. In this paper we briefly review the evolution of database systems over the last 4 decades and then focus on the most recent database developments discussing research and implementation challenges presented by modern database applications.
Content-based retrieval of compressed images
Gerald Schaefer
Department of Computer Science, Loughborough University, Loughborough, U.K.
Abstract.
Content-based image retrieval allows search for pictures in large image
databases without keyword or text annotations. Rather features are extracted
directly from the images and used as indices for search and retrieval. Much
progress has been made in deriving useful image features with most of these
features being extracted from (uncompressed) pixel data. However, the vast
majority of images today are stored in compressed form due to limitations in
terms of storage and bandwidth resources. This in turn means that for image
retrieval to be performed the images first need to be decompressed in order
to calculate image features, hence adding to the computational complexity of
the indexing and retrieval processes.
Addressing this issue leads to a different approach, namely that of
compressed-domain image retrieval. Here, image feature extraction, and hence
image retrieval, is performed directly in the compressed domain of images.
In general there are two approaches to compressed-domain image retrieval.
The first is based on existing compression techniques and tries to extract
useful information from the compressed data streams produced by these. The
second approach is to develop so-called 4-th criterion compression
algorithms where the data in compressed form is directly visually meaningful
and can hence be employed for image retrieval.
In my talk, I will present some compressed-domain image retrieval techniques
that we have developed over the past years. In particular, I will present a
method for retrieving images compressed by vector quantisation that uses
codebook information as image features. Retrieval of losslessly compressed
images obtained using lossless JPEG, can be retrieved using information
derived from the Huffman coding tables of the compressed files. Finally, I
will present CVPIC, a 4-th criterion image compression technique and show
that compressed-domain image retrieval based on CVPIC is not only able to
match the performance of common retrieval techniques on uncompressed images,
but even clearly outperforms these.
Chosen Problems of Designing Effective Multiple Classifier Systems
Michal Wozniak
Wroclaw University of Technology, Wroclaw, Poland
Abstract.
We encounter pattern recognition problems on an everyday basis. Therefore, methods of automatic pattern recognition form one of the main trends in Artificial Intelligence. The aim of each such recognition task is to classify a given object of interest by assigning it to some predefined category, on the basis of observing the features of the object. There is much current research into developing even more efficient and accurate recognition algorithms, like neural networks, statistical and symbolic learning to name only a few. Multiple classifier systems (MCSs) are currently the focus of intense research. In this conceptual approach, the main effort is concentrated on combining knowledge of the set of elementary classifiers.
There is a number of important issues while building the aforementioned MCSs. Firstly, how should classifiers be selected such that the decision making quality of the ensemble is superior to that of any individual classifier. This can be considered the problem of classifier synergy. So it seems interesting to select members of a committee with possibly different components. Another important issue is the choice of a collective decision making method. The first group of methods includes algorithms for classifier fusion at the level of their responses The second group of collective decision making methods exploit classifier fusion based on discriminant analysis, the main form of which are the posterior probability estimators, associated with probabilistic models of a given pattern recognition task. Design of new fusion classification models, especially those with a trained fuser block, are currently the focus of intense research.