Research Groups | Research Labs | Research Areas | Applied Research | Seminars

Content Start here

Databases

Databases support effective management and processing of a large amount of data, and have become an important topic in Computer Science. Traditional databases, however, are inadequate for numerous emerging applications (e.g., e-commerce, geographic information systems, advanced multimedia processing, etc.), where data and user- requirements are significantly different from those in conventional systems. Such applications demand the development of novel database techniques. The following are several topics that are being actively studied in the database community, and by the faculty members in our department.


Topic 1: XML (eXtensible Markup Language) is an important data representation for information exchange among different enterprises. Similar to HTML (which is the foundation of today's Internet), XML was originally designed as a language for writing structured documents. After several years of research, however, XML has evolved into a method for representing the content of a database. Several challenging issues exist towards implementing an "XML-database". For example, it is essential to develop solutions that can efficiently convert the tuples in a conventional database into the XML format (and vice versa), without losing the data semantics. Furthermore, query optimization directly on XML data is also a challenging problem.


Related website: ( homepage of Dr. Joseph Fong) http://www.cs.cityu.edu.hk/~jfong


Topic 2: Consider, for example, a set of two-dimensional points representing the locations of restaurants in Hong Kong. A range query would retrieve "all the restaurants that are within 1 mile from the Kowloon Tong subway station". Here, the search region is a circle that centers at the station, and its radius equals 1 mile. A nearest neighbor query, on the other hand, would identify "the restaurant that is closest to the Star Ferry". In many cases, a user may be interested in statistical results, as in "find the number of vehicles currently near the International Finance Center" (as opposed to the details of these vehicles), or "what is the average distance between a hotel to its nearest restaurant?" (i.e., the result includes a single value). A spatial database efficiently supports the above operations.


Related website: (homepage of Dr. Yufei Tao) http://www.cs.cityu.edu.hk/~taoyf


Topic 3: Multimedia databases aim at fast retrieval of multimedia objects (e.g., images, videos, etc.) according to various query predicates. For example, "find the shots in a movie that contain McDonalds", "retrieve the images that are most similar to a given image", or "identify those pictures that are about sun-sets". Our ongoing research considers (i) the application of object-oriented modeling to preprocess multimedia objects for constructing object-relational data warehouses, (ii) complex object retrieval using join indexes, and (iii) semantic modeling through media views, and query re-writing.


Related website: (homepage of Dr. Qing Li) http://www.cs.cityu.edu.hk/~csqli/

| [Goto TOP]