The Experts below are selected from a list of 27 Experts worldwide ranked by ideXlab platform
Gerhard Rigoll - One of the best experts on this subject based on the ideXlab platform.
-
multimodal meeting analysis by segmentation and classification of meeting events based on a Higher Level Semantic approach
International Conference on Acoustics Speech and Signal Processing, 2005Co-Authors: S Reiter, Sascha Schreiber, Gerhard RigollAbstract:This paper encompasses the analysis of meetings for segmentation into sub-genres. Therefore, an approach on a Higher Semantic Level has been chosen. The algorithms make use of the results of specialized recognizers like a speaker turn detector and a gesture recognizer. Basically, the goal of this investigation was to answer the question, how well meeting analysis is possible if only the results of these recognizers are available. After introducing briefly the basics of these recognizers, two slightly different methods for the segmentation are presented. The results show the potential of the used methods to find the segment boundaries and to categorize the detected segments into sub-genres (also called meeting events or group actions). Based on this segmentation, further analysis regarding topic detection and content extraction can be accomplished.
S Reiter - One of the best experts on this subject based on the ideXlab platform.
-
multimodal meeting analysis by segmentation and classification of meeting events based on a Higher Level Semantic approach
International Conference on Acoustics Speech and Signal Processing, 2005Co-Authors: S Reiter, Sascha Schreiber, Gerhard RigollAbstract:This paper encompasses the analysis of meetings for segmentation into sub-genres. Therefore, an approach on a Higher Semantic Level has been chosen. The algorithms make use of the results of specialized recognizers like a speaker turn detector and a gesture recognizer. Basically, the goal of this investigation was to answer the question, how well meeting analysis is possible if only the results of these recognizers are available. After introducing briefly the basics of these recognizers, two slightly different methods for the segmentation are presented. The results show the potential of the used methods to find the segment boundaries and to categorize the detected segments into sub-genres (also called meeting events or group actions). Based on this segmentation, further analysis regarding topic detection and content extraction can be accomplished.
Sascha Schreiber - One of the best experts on this subject based on the ideXlab platform.
-
multimodal meeting analysis by segmentation and classification of meeting events based on a Higher Level Semantic approach
International Conference on Acoustics Speech and Signal Processing, 2005Co-Authors: S Reiter, Sascha Schreiber, Gerhard RigollAbstract:This paper encompasses the analysis of meetings for segmentation into sub-genres. Therefore, an approach on a Higher Semantic Level has been chosen. The algorithms make use of the results of specialized recognizers like a speaker turn detector and a gesture recognizer. Basically, the goal of this investigation was to answer the question, how well meeting analysis is possible if only the results of these recognizers are available. After introducing briefly the basics of these recognizers, two slightly different methods for the segmentation are presented. The results show the potential of the used methods to find the segment boundaries and to categorize the detected segments into sub-genres (also called meeting events or group actions). Based on this segmentation, further analysis regarding topic detection and content extraction can be accomplished.
Krzysztof J Kochut - One of the best experts on this subject based on the ideXlab platform.
-
Semantic enhancement engine a modular document enhancement platform for Semantic applications over heterogeneous content
2002Co-Authors: Brian Hammond, Amit P Sheth, Krzysztof J KochutAbstract:Traditionally, automatic classification and metadata extraction have been performed in isolation, usually on unformatted text. SCORE Enhancement Engine (SEE) is a component of a Semantic Web technology called the Semantic Content Organization and Retrieval Engine (SCORE). SEE takes the next natural steps by supporting heterogeneous content (not only unformatted text), as well as following up automatic classification with extraction of contextually relevant, domain-specific (i.e., Semantic) metadata. Extraction of Semantic metadata not only includes identification of relevant entities but also relationships within the context of relevant ontology. This paper describes SEE's architecture, which provides a common API for heterogeneous document processing, with discrete, reusable and highly configurable modular components. This results in exceptional flexibility, extensibility and performance. Referred to as SEE modules (SEEMs), which are divided along functional lines, these processors perform one of the following roles: restriction (determine the segments of the input text to operate upon); enhancement (discover textual features of Semantic interest); filtering (augment, remove or supplement the features recognized); or outputting (generate reports, annotate the original, update databases, or other actions). Each SEEM manages its configuration options and is arranged serially in virtual pipelines to perform designated Semantic tasks. These configurations can be saved and reloaded on a per-document basis. This allows a single SEE installation to act logically as any number of Semantic Applications, and to compose these Semantic Applications as needed to perform even more complex Semantic tasks. SEE leverages SCORE's unique approach of creating and using large knowledge base in Semantic processing. It enables SCORE to provide flexible handling of highly heterogeneous content (including raw text, HTML, XML and documents of various formats); reliable automatic classification of documents; accurate extraction of Semantic, domain-specific metadata; and extensive management of the enhancement processes including various reporting and Semantic annotation mechanisms. This results in SCORE's advanced capability in heterogeneous content integration at a Higher Semantic Level, rather than syntactical and structural Level approaches based on XML and RDF, by supporting and exploiting domain specific ontologies. This work also presents an approach to automatic Semantic annotation, a key scalability challenge faced in realizing the Semantic Web.
Joemon M Jose - One of the best experts on this subject based on the ideXlab platform.
-
adaptive image retrieval using a graph model for Semantic feature integration
Multimedia Information Retrieval, 2006Co-Authors: Jana Urban, Joemon M JoseAbstract:The variety of features available to represent multimedia data constitutes a rich pool of information. However, the plethora of data poses a challenge in terms of feature selection and integration for effective retrieval. Moreover, to further improve effectiveness, the retrieval model should ideally incorporate context-dependent feature representations to allow for retrieval on a Higher Semantic Level. In this paper we present a retrieval model and learning framework for the purpose of interactive information retrieval. We describe how Semantic relations between multimedia objects based on user interaction can be learnt and then integrated with visual and textual features into a unified framework. The framework models both feature similarities and Semantic relations in a single graph. Querying in this model is implemented using the theory of random walks. In addition, we present ideas to implement short-term learning from relevance feedback. Systematic experimental results validate the effectiveness of the proposed approach for image retrieval. However, the model is not restricted to the image domain and could easily be employed for retrieving multimedia data (and even a combination of different domains, eg images, audio and text documents).