diploma thesis

Facet-based Personal Media Management

Task

There is no doubt about the tremendous growth of personal digitized data including e-mails, conversation histories, blogs, but also health care, insurance and banking documents as well as an ever growing number of media documents including commercial music, movies as well as personal photos, videos etc. The efficient home-based management of personal media using intuitive and accessible interfaces for everyday users will be a grand challenge of future systems. Recent research has shown that browsing media documents according to their multi-dimensional meta data facets has proven to be a promising development. Thereby the huge personal document sets are filtered according to several dimensions such as time, geographic location, involved persons, tags, rating etc. Since a large quantity of data is available on each dimension, the dimensional complexity should be reduced by means of powerful filtering. This work should investigate how to combine filter dimensions in a visual way while maintaining an overview and displaying contextual information. To overcome the standard way of displaying result data sets in long lists as currently provided by search engines, the usage of several structural visualizations and seamless transitions between them shall be investigated. Several questions need to be addressed by this work, among them: Which typical and generalizable filter dimensions or meta data facets should be used for personal information and media management? How should result sets be ordered according to these dimensions? What alternatives are available for displaying resulting documents as opposed to long item lists or matrices? How to transparently visualize the influence of filters in the result set of documents? How can filters be applied in a sequential as well as parallel fashion without overburdening the user? How to provide history functionality in such a system? How can users maintain their data, add new tags, ratings or filter details? Is there still a separation between interface parts and result visualizations? The following results are expected from this work:

Analysis of use cases and requirements regarding the potential users, their tasks and system environments as well as the media documents involved.
Choice & description of media document types and meta data suitable for generalization.
Careful analysis of the state-of-the-art and related research in the field of facet-based media managers with regard to unsolved problems, current systems, target users and platforms
- including appropriate information visualization techniques, e.g. 2.5/3D solutions.
Conceptual design of specific novel approaches to support facet-based media management
- by selecting an appropriate and focused application scenario along with the data,
- by designing visual filters, their transitions, data couplings and result visualizations,
- by taking authoring tasks (e.g. structuring, tagging, rating) into account.
- Prototypical realization of selected concepts or parts of a future media management system and appropriate testing depending on the results, e.g. by usability tests.

Further information

Author(s):	Markus Weiland
Duration:	07/10/2007 - 01/10/2008
University teacher:	Prof. Dr.-Ing. Klaus Meißner
Supervisor(s):	Dr.-Ing. Raimund Dachselt
Institute, Chair:	SMT, Multimedia Technology
Related Projects:	INPERIC