Werner Klieber

Evaluating a System for Interactive Exploration of Large, Hierarchically Structured Document Repositories

The InfoSky visual explorer is a system enabling users to interactively explore large, hierarchically structured document collections. Similar to a real-world telescope, InfoSky employs a planar graphical representation with variable magnification. Documents of similar content are placed close to each other and displayed as stars, while collections of documents at a particular level in the hierarchy are visualised as bounding polygons. Usability testing of an early prototype implementation of InfoSky revealed several design issues which prevented users from fully exploiting the power of the visual metaphor. Evaluation results have been incorporated into an advanced prototype, and another usability test has been conducted. A comparison of test results demonstrates enhanced system performance and points out promising directions for further work

Content-based Video Retrieval and Summarization using MPEG-7

Retrieval in current multimedia databases is usually limited to browsing and searching based on low-level visual features and explicit textual descriptors. Semantic aspects of visual information are mainly described in full text attributes or mapped onto specialized, application specific description schemes. Result lists of queries are commonly represented by textual descriptions and single key frames. This approach is valid for text documents and images, but is often insufficient to represent video content in a meaningful way. In this paper we present a multimedia retrieval framework focusing on video objects, which fully relies on the MPEG-7 standard as information base. It provides a content-based retrieval interface which uses hierarchical content-based video summaries to allow for quick viewing and browsing through search results even on bandwidth limited Web applications. Additionally semantic meaning about video content can be annotated based on domain specific ontologies, enabling a more targeted search for content. Our experiences and results with these techniques will be discussed in this paper.

Visual Knowledge Discovery in Dynamic Enterprise Text Repositories

Knowledge discovery involves data driven processes where data is transformed and processed by various algorithms to identify new knowledge. KnowMiner is a service oriented framework providing a rich set of knowledge discovery functionalities with focus on text data sets. Complementing results of automatic machine analysis with the immense processing power of human visual apparatus has the potential of significantly improving the process of acquiring new knowledge. VisTools is a lightweight visual analytics framework based on multiple coordinated views (MCV) paradigm designed for deployment atop the KnowMiner’s service architecture. In this paper we briefly present both frameworks and, driven by real-world customer requirements, describe how visual techniques can be synergistically combined with machine processing for effective analysis of dynamically changing, metadata-rich text documents sets.

Plagiarism Detection in Large Sets of Press Agency News Articles

The advent of electronic media enabling rapid publishing and instant access to news articles has vastly increased the working pace of the media industry. Press agencies can no longer afford to neglect plagiarisms formerly perceived as irrelevant because such acts instantly void the commercial value of original and potentially exclusive news articles. From a different point of view, individuals and organizations of public interest require instant evaluation of impact of actions taken and diffusion of statements made in the media. This publication presents a methodology and a prototypical system for analyzing plagiarisms and, in general, diffusion of textual content in newspaper articles and reports on preliminary application results obtained in the context of Austrias press agency APA

A Visual Query Interface for a Very Large Newspaper Article Repository

The archives of large national and international news agencies typically contain millions of articles featuring significant textual content and annotated metadata. Boolean queries and relevance ranked result lists have been traditional means of inquiry in such a context. This article presents an interface for query formulation and visual query analysis for the news article repository of the Austrian Press Agency, combining traditional means of inquiry with several visualisation components. All parts of the interface are fully synchronised, allowing users to employ the power of visual metaphors for analysis and refinement while retaining the simplicity of conventional query formulation

Knowminer Search - A Multi-visualisation Collaborative Approach to Search Result Analysis

The amount of information available on the internet and within enterprises has reached an incredible dimension. Efficiently finding and understanding information and thereby saving resources remains one of the major challenges in our daily work. Powerful text analysis methods, a scalable faceted retrieval engine and a well-designed interactive user interface are required to address the problem. Besides providing means for drilling-down to the relevant piece of information, a part of the challenge arises from the need of analysing and visualising data to discover relationships and correlations, gain an overview of data distributions and unveil trends. Visual interfaces leverage the enormous bandwidth of the human visual system to support pattern discovery in large amounts of data. Our Know miner search builds upon the well-known faceted search approach which is extended with interactive visualisations allowing users to analyse different aspects of the result set. Additionally, our system provides functionality for organising interesting search results into portfolios, and also supports social features for rating and boosting search results and for sharing and annotating portfolios.

Cross Media Retrieval in Knowledge Discovery

Recent trends show that more and more digital cameras, video cameras and DVD recorders are sold and the number of emails and other messages sent increases each year. For example it is estimated that there will be nearly 300 million digital image capture devices in use worldwide through 2004, capturing about 29 billion digital pictures [12]. Users not only produce huge amounts of content, but this content is also spread over many different media types and document formats. Storing all contents implicates a great extent of documents in large, heterogeneous data repositories. Performing effective retrieval in heterogeneous repositories requires new approaches: documents in different formats containing different media types have to be transformed to a common denominator, and relations between contents of different media types must be established. This allows direct comparison of contents of different media types. In this paper we present a prototype called Magick that implements such an approach to cross-media retrieval.