IST LIVE

Content related to "annotation-principles"

(1) Sport Database metadata retrieval scenarios

In this document we suppose that the content domain is limited to the sport domain. We only use terms available and defined by the SportML schema. We propose that the first demon-strator capabilities are limited to the sport domain.

(1) The Role of MPEG-7 in Semantic Annotation and the Cross-Media Publishing Process

During the development of a knowledge-based audio-visual information system the authors of this article defined a conceptual system architecture based on MPEG-7 as the general description scheme for the media assets in the middleware. This concept was not only used to achieve a high abstraction and independence of the underlying media asset management system, it was also and primarily used as the basis of a semantic indexing process. Based on lightweight ontologies the descriptions of the media assets were associated with semantic concepts. Semantically annotated MPEG-7 assets were then propagated to the presentation layer, thus allowing the implementation of a variety of publication scenarios, including cross-media scenarios for the creation of concise video summaries.

(1) Knowledge Acquisition from Multimedia Content

Proceedings of the First International Workshop, KAMC 2007 Genova, Italy, December 5, 2007. In recent years significant advances have been made in the area of automatic ex- traction of low-level features from audiovisual content. However, little progress has been achieved in the identification of high-level semantic features or the effective combination of semantic features derived from different modalities. Knowledge acquisition is becoming a key-enabling factor of the above tasks towards more scalable and reliable solutions, and thus its automation is becoming critical. As the deployment of knowledge enhances the robustness of extraction while on the other hand the continuous extraction of semantic information can enrich this knowledge, synergistic approaches that combine multimedia extraction and knowledge evolution in a bootstrapping common framework introduce new opportunities in semantic multimedia applications. Integration with additional sources of information, e.g. by using human annotation tools or real-time event services, may further simplify and disambiguate semantic multimedia information systems. Moreover, adaptation to a particular domain, for example to sports events, such as the Olympic games, is essential in order to reduce the complexity of multimedia analysis. In this context, unified modelling and representation of multimedia and domain-specific knowledge, ontology evolution, and standard and non-standard inference services for multimodal semantic knowledge fusion, form cutting edge technologies. The aim of this workshop is to intensify the exchange of ideas between the different research communities involved which range from multimedia analysis to reasoning with ontologies. The submitted contributions published in these proceedings therefore reflect current research in this area: the topics range from multimedia classification based on textual information, content based shot classification, feature extraction to image classification based on ontologies. The submitted papers cover different application domains, i.e. broadcasted news or legal documents. We would like to thank all members of the program committee for supporting us in the reviewing process, the organizers of the main conference SAMT 2007 to which this workshop was co-located - especially Yannis Avrithis, Michela Spagnuolu and Francesco Robbiano - for their kind support throughout the organizational process.We also would like to thank the authors for their willingness to revise their initial submissions based on the reviewers comments. Finally we would like to thank our invited speakers, Fabio Ciravegna and Alan Smeaton for their willingness to give a talk at our workshop.

(1) D5.2 Report On Live Human Annotation

This document reports on human annotation within the LIVE project. First it gives an overview about different annotation types that are useful for the LIVE staging of media events. It then summarizes the requirements for manual annotation by collecting results from potential users, e.g. from discussions performed with broadcasters, reporters, editors and video jockeys (VJs). It defines the necessary content metadata needed within the LIVE system, gives an overview over existing tools and describes the tools developed for the LIVE project. Finally, user evaluations of the developed tools that were performed with professional users from the ORF are described at the end of the document.

(1) Mind the Gap - Requirements for the combination of content and knowledge

Semantic enrichment of content can be done manually, which is expensive, or automatically, which is error-prone. In particular, automatic semantic enrichment must be aware of the gap between the semantics that are directly retrievable from the content and those which can be inferred within a given interpretative context. We report on a model for content and knowledge which distinguishes between three descriptive levels: information relating directly to the resource, to the metadata of the resource and to the subject matter addressed by the content. This model addresses five fundamental requirements for automation: formality, interoperability, multiple interpretations, contextualization, and independence of knowledge items from the resource’s content.

(1) Description of Online and Offline Metadata Extraction out of Sports Videos

We focus on online and offline metadata extraction and annotation out of sports videos. The main benefit of our method is immediate and automatic extraction and annotation of metadata by giving semantics to combinations of heterogeneous low-level visual features. It brings new opportunities for efficient utilisation of sports video in improved ways, and is easily customized to address the characteristics. Firstly, semantic scene classification is described, including key-frames extraction, similarities determination between shots, and rule based estimation of scene boundaries. Secondly, fuzzy logic based categorizing is presented, including paradigm, Fuzzy membership function, and fuzzy feature generation and similarity measure. Thirdly, automatic sports video annotation is proposed, including robust dominant colour region detection, combined motion feature analysis. This work has been evaluated in the TRECVID 2007 competition.