MANAGING AND EXPLORING MEDIA USING SEMANTIC REGIONS: A SPATIAL INTERFACE SUPPORTING USER-DEFINED MENTAL MODELS

Abstract

Computer users deal with large numbers of personal media objects such as images, audio clips, voice mails, video clips, web pages, emails, and various document files. Users often struggle to interpret, explore, arrange, and use personal media objects because of three major problems; an ever increasing amount of personal media data, rigid organizing metaphor, and difficulty in rapid data access.

With the progress of computer hardware and digitization technologies at the rate of Moore's law, users may generate, acquire, and store more and more personal media data on their personal machines. However, the means available for users to organize and customize their personal media information spaces are extremely poor and driven mostly by storage and distribution models, not users' needs. Furthermore, this rigid and system-oriented metaphor causes wide and deep file folder hierarchies and often forbids users to rapid data access from storage.

This dissertation uses a graphical mechanism for spatially organizing personal media data based on users' mental models and introduces an innovative interaction metaphor to apply users' mental models to personal media data. Semantic Regions introduces an innovative way to construct users' mental models by drawing regions on 2D space and specifying the semantics for each region so that users can apply various personal ontologies to personal media data using the fling-and-flock metaphor. The prototype application, MediaFinder, validates the usability of the interface, particularly in comparison with alternative approaches. Contributions of this dissertation include:

- Semantic Regions provides a formal model of spatial and dynamic reorganization of personal media data based on users' mental models. It extends the concept from a system-oriented file management system to a user-oriented personal media management system by employing the semantics of personal media data.

- The MediaFinder personal media management tool, which uses Semantic Regions, the fling-and-flock metaphor, and flexible categorization capabilities to explore and manage personal media data sets. MediaFinder's object-oriented architecture can easily be extended to support variants of the Semantic Regions model.

- The design and implementation of interaction models that support the personal media management tasks such as organization, meaning extraction, navigation, search, indexing, and navigation.

- Two usability studies provided preliminary insight into the utility of Semantic Regions and led to design improvements for the construction and operation of Semantic Regions.

- A framework for extending the Semantic Regions model, including the descriptions of possible extensions.