|
Bildannotation im semantischen Web
by Andreas Walter, Daniel Koch, ix Magazin
source: IX Magazin, Ausgabe 11, 2008, Seiten 126-130
ABSTRACT
Evaluation of the navigation through image parts in the ImageNotion application
by Andreas Walter, Gabor Nagypal, Simone Braun
Poster in Proceedings of OnTheMove Federated Conferences 2008 (DAO, COOP, GADA, ODBASE), Monterrey, Mexico, Lecture Notes in Computer Science, Springer, 2008
ABSTRACT
In our work on the ImageNotion methodology and tools, we apply semantic technologies on image archives. In this paper, we show evaluation results on our work on the user interface for semantic image search and expected navigation through image parts. We conducted an online survey with more than
hundred participants. A unique feature of our evaluation is that our evaluators filled the survey based on a concrete, working semantic application, i.e., based on the publicly available online version of our system.
Evaluation of the navigation through image parts in the ImageNotion application
by Andreas Walter, Gabor Nagypal, Simone Braun
source: Poster in Proceedings of OnTheMove Federated Conferences 2008 (DAO, COOP, GADA, ODBASE), Monterrey, Mexico, Lecture Notes in Computer Science, Springer, 2008
br />ABSTRACT
In our work on the ImageNotion methodology and tools, we apply
semantic technologies on image archives. In this paper, we show evaluation results
on our work on the user interface for semantic image search and expected
navigation through image parts. We conducted an online survey with more than
hundred participants. A unique feature of our evaluation is that our evaluators
filled the survey based on a concrete, working semantic application, i.e., based on
the publicly available online version of our system.
Evaluation of Semantic Search Using The ImageNotion Application
by Andreas Walter, Gabor Nagypal, Khaled Nagi
Proceedings of the IADIS International Conference WWW/Internet 2008, Freiburg, Germany, pp. 442-447
ABSTRACT
Within our ImageNotion application, we develop and combine these techniques to improve end user experience by providing innovative query refinement and navigation features.
We have already reported on our work on the ImageNotion methodology and tools [6] which includes collaborative and work-integrated ontology creation, semantic annotation, and the possibility to automate the annotation process using text mining and image processing techniques to construct the image archive. In this paper, we describe some novel techniques in our ImageNotion application for visual query refinement and navigation through image parts to be used by image
searchers. We conducted an online survey with more than hundred participants who tested the publicly available online version of this system. We analyze the results, and show what users think about the potential of semantic techniques for the exploration of image archives based on this concrete and working semantic image search application.
The insights we gain will guide the further development of the ImageNotion system, and are interesting for all kinds of semantic multimedia information systems in general.
Concept Detection and Keyframe Extraction Using a Visual Thesaurus
by Evaggelos Spyrou, Giorgos Tolias, Phivos Mylonas, Yannis Avrithis
ABSTRACT
This paper presents a video analysis and indexing approach based on concept detection and
keyframe extraction employing a visual thesaurus image representation. Color and texture descriptors
are extracted from coarse regions of each video frame and a visual thesaurus constructed by clustering
regions in the descriptor space. The clusters, called region types, are used as basis for representing local
material information in frames through the construction of a model vector for each frame, which re
ects the composition of the visual scene in terms of region types. Model vector representation is used for
keyframe selection either in each video shot or across an entire sequence; the selection process ensures
that all region types are represented, while the selected keyframes illustrate diversity in terms of model
vectors, i.e., global visual content. A number of high-level concept detectors is then trained using the
same model vector representation and a global keyframe annotation. To enhance detection performance
per shot, detection is employed on the selected keyframes of each shot, and a framework is proposed
for working on very large data sets. Finally, detection performance is investigated after applying Latent
Semantic Analysis to model vectors, exploiting latent relations among region types.
Using Visual Context and Region Semantics
for High-Level Concept Detection
by Phivos Mylonas, Student Member, IEEE, Evaggelos Spyrou, Student Member, IEEE,
Yannis Avrithis, Member, IEEE and Stefanos Kollias, Member, IEEE
TRANSACTIONS ON MULTIMEDIA, APRIL 2008
ABSTRACT
In this paper we investigate detection of high-level concepts in multimedia content through an
integrated approach of visual thesaurus analysis and visual context. In the former, detection is based
on model vectors that represent image composition in terms of region types, obtained through clustering
over a large data set. The latter deals with two aspects, namely high-level concepts and region types of the
thesaurus, employing a model of a priori specified semantic relations among concepts and automatically
extracted topological relations among region types; thus it combines both conceptual and topological
context. A set of algorithms is presented, which modify either the confidence values of detected concepts,
or the model vectors based on which detection is performed. Visual context exploitation is evaluated on
TRECVID and Corel data sets and compared to a number of related visual thesaurus approaches.
The Imagination project: Image-based Navigation in Multimedia Archives
by Andreas Walter, Gabor Nagypal
Proceedings of the IADIS International Conference WWW/Internet 2008, Freiburg, Germany, pp. 251-258
ABSTRACT
The IMAGINATION project provides image-based navigation for digital cultural and scientific resources. Users can click on parts of an image to find other, interesting images to a given context. In this paper, we give an overview on the components of the IMAGINATION project. To allow the navigation through images, high quality semantic metadata are generated automatically. Therefore it combines automated processes for person and object detection, face detection and
identification in images together with text mining techniques that exploit domain specific ontologies.
The Utility of Specific Relations in Ontologies for Image Archives
by Andreas Walter, Gabor Nagypal
Proceedings of the third International Conference on Semantic Systems (I-Semantics), 3-5 September 2008, Graz, Austria, pp. 17-24
ABSTRACT
The ImageNotion methodology and tools [Walter and Nagypal (2007)] support collaborative ontology creation and semantic image annotation in one integrated web-based application. In previous evaluations, we received very positive feedback from our users about this visual methodology. Users found the approach understandable and useful. So far, the ImageNotion methodology supports for the sake of simplicity only three kinds of relations: broader, narrower and unnamed relations. We were interested, however, whether users would find it useful to have more kinds of relations, which would also make our ontology more expressive. We therefore evaluated in an online survey what users think of this issue. The evaluation was based on the publicly available online prototype of the system. We could attract more than one hundred participants. This paper analyzes the results of this survey.
Evaluating semantic techniques for the exploration of image archives on the example of the ImageNotion system
by Andreas Walter, Gabor Nagypal, Khaled Nagi
source: Alexandria Engineering Journal (AEJ), Vol. 47, No. 4. ISSN: 1110-0168 Springer, pp. 327-338
ABSTRACT
Semantic technologies can provide better search results in image archives than full-text search.
Semantic technologies require background knowledge stored in ontologies and semantically annotated images. Therefore, a methodology for ontology creation and image annotation is required that is usable in practice also for annotators with minimal expertise of semantic technologies.
Therefore, we created and previously described the ImageNotion methodology and application.
In this paper, we are exploring possible applications of semantic technologies to improve search results and user friendliness of the navigation through image archives. We executed an online survey with more than hundred participants who tested the online version of the ImageNotion application. In this paper, we analyze the results, and show what users think about the potential of semantic technologies based on a concrete, working image search application.
ImageNotion as a mashup service for a semantic image web
by Andreas Walter, Gabor Nagypal
International Workshop on Interacting with Multimedia Content in the Social Semantic Web (IMC-SSW), Proceedings of the 3rd International Conference on Semantic and Digital Media Technologies (SAMT 2008)
ABSTRACT
Most web image archives still use plain text to annotate images. The ImageNotion system extends the state-of-the-art by providing semantic annotation of images and their parts. In this paper, we show how to implement a mashup of ImageNotion with popular web image archives such as Flickr.
This allows our users to load the images from external image platforms in ImageNotion and to automatically create semantic annotations using ImageNotion.
Further, they can use the advanced search features of ImageNotion on those images, including the search for related images. In addition, we also show how to extend this mashup to a semantic web service. Such a service creates semantic image annotations semi-automatically, and thus makes those images available for
processing on the Semantic Web.
PRIMO - Towards Privacy Aware Image Sharing
by Andreas Walter, Thorben Burghardt, Erik Buchmann, Klemens Böhm
2nd Workshop on Collective Intelligence in Semantic Web and Social Networks (CISWSN 2008), Proceedings of the IEEE/WIC/ACM Joint International Conference on Web Intelligence and Intelligent Agent Technology 2008, Sydney, Australia
ABSTRACT
A growing number of users in Web 2.0 based social network sites and photo sharing portals upload millions of images per day. In many cases, this leads to serious privacy threats. The images reveal not only the personal relationships and attitudes of the uploader, but of other persons displayed in the images as well. In this paper, we propose the
PRIMO system architecture for privacy-aware image sharing.
Our approach is based on semantic annotations, face recognition and user-defined privacy rules. PRIMO connects to many social network sites and photo sharing portals via the OpenSocial API and proprietary interfaces.
D15 User Guide of ImageNotion (pdf file) - 3MB (public project deliverable)
SUMARY This is the end user guide of the ImageNotion application. The ImageNotion application represents the integrated result of the IMAGINATION project that is visible for the end users, and thus for the general public. D14 IMAGINATION Ontology (pdf file) - 700 KB (public project deliverable)
ABSTRACT
This document gives an overview of the ImageNotion collaborative ontology development methodology that is used in the IMAGINATION project for ontology development. ImageNotion defines a methodology where ontology development is embedded into the workflow of semantic image annotation. The ImageNotion web application supports this methodology. The document also contains a snapshot of the IMAGINATION ontology in OWL/RDF format. Efficient Integration of Semantic Technologies for professional Image Annotation and Search (pdf file) - 0,4 MB by Andreas Walter and Gabor Nagypal. Proceedings of the IADIS international conference e-Society 2008. ABSTRACT
Professional image annotation and search concerns two user groups. The first group consists of image agencies who aim to sell their images. The second group is formed by image buyers who search and use these images for the illustration of printed materials or for advertising. Professional image annotation and search is based on text based techniques for many years. Semantic technologies help improve both, the quality of the image annotation for image editors and the possibilities for the navigation through an image archive for image searcher. In this paper, we present the overall result of our research. We first collected the requirements of the two user groups for an effective usage and integration of semantic technologies in image archives. Based on these requirements, we created a visual methodology and application which fulfils these needs.
Digitalbilder mit elektronischen Etiketten eindeutig beschreiben (pdf file) 1MB (also in doc file) by Andreas Walter Annual Report of FZI. ABSTRACT Die Software ImageNotion unterstützt Bildredakteure bei der Beschreibung von Bildmotiven durch einfach anzulegende elektronische Etiketten, die, einmal erzeugt, immer wieder verwendet werden können. Bildsuchmaschinen erkennen daran, was auf einem Bild zu sehen ist. Die kollaborative Technik zur Bilddokumentation wurde am FZI entwickelt.
Von Tags zu semantischen Beziehungen: kollaborative Ontologiereifung by Simone Braun, Andreas Schmidt, Andreas Walter, Valentin Zacharias. In: Birgit Gaiser and Stefanie Panke (eds.): Good Tags and Bad Tags - Workshop „Social Tagging in der Wissensorganisation“, 2008
Regions of interest for accurate object detection (pdf file) - 0,6 MB by P. Kapsalas, K. Rapantzikos, A. Sofou, Y. Avrithis
Proc. of Sixth International Workshop on Content-Based Multimedia Indexing (CBMI 2008), 18-20th June, 2008, London, UK ABSTRACT
In this paper we propose an object detection approach that extracts a limited number of candidate local regions to guide the detection process. The basic idea of the approach is that object location can be determined by clustering points of interest and hierarchically forming candidate regions according to similarity and spatial proximity predicates.
Statistical validation shows that the method is robust across a substantial range of content diversity while its response seems to be comparable to other state of the art object detectors.
The combination of techniques for automatic semantic image annotation generation in the IMAGINATION application (pdf file) by Andreas Walter and Gabor Nagypal. 5th European Semantic Web Conference, ESWC 2008, Tenerife, Canary Islands, Spain, June 1-5, 2008 Proceedings. ABSTRACT The IMAGINATION project provides image-based navigation for digital cultural and scientific resources. Users can click on parts of an image to find other, interesting images to a given context. In this paper, we present the core parts of the IMAGINATION application. To allow the navigation through images, this application automatically generates high quality semantic metadata. Therefore it combines automated processes for person and object detection, face detection and identification in images together with text mining techniques that exploit domain specific ontologies.
ImageNotion - Methodology, Tool Support and Evaluation (pdf file) by Andreas Walter and Gabor Nagypal. ODBASE 2007 27. - 29. November 2007, Algave, Portugal ABSTRACT The content of image archives changes rapidly. This makes the traditional separation of ontology development and image annotation steps no longer feasible. In this paper, we present an approach, lermed ImageNotion that allows for the collaborative development of domain ontologies directly by
domain experts with minimal ontology experience. ImageNotion is both a methodology based on the idea of the ontology maturing model, and the name of the tool supporting this methodology. ImageNotion embeds the creation of ontology entities, termed imagenotions, into the work process of creating semantic annotations of images and their parts. Both the creation of imagenotions and the creation of image annotations are visual, user friendly processes, implemented by a web application that integrates all of the required functionality in one consistent framework. Besides the theoretical concepts, this paper also presents the results of our evaluation of the system with experienced image annotators and librarians having minimal ontology background.
The utility of specific relations in ontologies for image archives
(pdf file) - 0,7 MB
by Andreas Walter and Gábor Nagypál.
ABSTRACT
The ImageNotion methodology and tools support collaborative ontology creation and semantic image annotation in one integrated web-based application. In previous evaluations, we received very positive feedback from our users about this visual methodology. Users found the approach understandable and useful. So far, the ImageNotion methodology supports for the sake of simplicity only three kinds of relations: broader, narrower and unnamed relations. We were interested, however, whether users would find it useful to have more kinds of relations, which would also make our ontology more expressive. We therefore evaluated in an online survey what users think of this issue. The evaluation was based on the publicly available online prototype of the system. We could attract more than one hundred participants. This paper analyzes the results of this survey. D12 Report on the object detection algorithms (pdf file) - 2MB (public project deliverable)
ABSTRACT
This deliverable assesses various approaches found in the scientific literature. Our algorithms are presented in detail and their performance is measured. Plans for future work are given.
The role of object detection within the Imagination project is to enable an image to annotate itself.
D11 Report on Image Recognition Algorithms (pdf file) - 500 KB (public project deliverable) ABSTRACT It deals with the algorithms developed in the field of image recognition.
It covers especially the algorithms for face detection and for face recognition. Face detection and identification are playing a significant part within the imagination project in order to automate the process of annotation and the creation of knowledge. These tools are combined with person segmentation tools and text mining tools in order to enhance metadata information for images either they have or don't have textual
captions.
D10 Report on the IMAGINATION Text Mining algorithms (pdf file) - 700 KB (public project deliverable)
ABSTRACT
This deliverable proposes text mining algorithms to be used in IMAGINATION, based on the state-of-the-art analysis given in IMAGINATION deliverable D3 and the
requirements described in IMAGINATION deliverable D6.
IMAGENOTION - Collaborative Semantic Annotation of Images and Image Parts and Work Integrated Creation of Ontologies (pdf file) - 220 KB by Andreas Walter and Gábor Nagypál.
accepted publication for the SABRE Conference on Social Semantic Web (CSSW 2007)
24. - 27. September 2007, Leipzig, Germany
ABSTRACT
The paper presents the imagenotion tool, that allows both – the semantic annotation of images and image parts together with the maturing of ontologies by using our so called imagenotion formalism in a work integrated environment.
Ontology Maturing: a Collaborative Web 2.0 Approach to Ontology Engineering
(pdf file) - 400 KB
by Simone Braun, Andreas Schmidt, Andreas Walter, Gábor Nagypál and Valentin Zacharias.
Proceedings of the Workshop on Social and Collaborative Construction of Structured Knowledge (CKC), 16th International World Wide Web Conference (WWW 2007), Banff, Alberata, Canada.
ABSTRACT
The proposed ontology maturing processes is based on the insight that ontology engineering is a collaborative informal learning process and for which we analyze characteristic evolution steps
and triggers that have users engage in ontology engineering within their everyday work processes. This model integrates tagging and folksonomies with formal ontologies and shows maturing pathways between them. As implementations of this model, we present two case studies and the corresponding tools. The first is about image-based ontology engineering introducing so-called imagenotions), the second about ontology-enabled social bookmarking (SOBOLEO). Both of them are inspired by lightweight Web 2.0 approaches and allow for realtime collaboration.
Semantical interoperability with IMAGINATION content using standardized ontologies (pdf file) - 350 KB
by Andreas Walter and Gábor Nagypál (2007):
invited paper to the DELOS-MultiMatch workshop on "Ontology-driven Interoperability", Pisa, Italy (15 February 2007).
ABSTRACT
The IMAGINATION project provides image-based navigation for digital cultural and scientific resources. Users can click on parts of an image to find other, interesting images to a given context. The combined application of object detection and identification in images together with text mining
techniques exploiting domain specific ontologies will help generate highquality semantic metadata. We want to share this metadata with other information systems, e.g. in the domain of cultural heritage. This paper describes the requirements of IMAGINATION that must be fulfilled to reach this goal and analyzes CIDOC-CRM, FRBR and MPEG-7 based on these requirements.
Ontology supported automatic generation of high-quality semantic metadata (pdf file) - 250 KB
by Ümit Yoldas and Gábor Nagypál.
Proceedings of the On the Move to Meaningful Internet Systems 2006: CoopIS, DOA, GADA, and ODBASE, Springer LNCS 4275, pp. 791-806. ABSTRACT Large amounts of data in modern information systems, such as the World Wide Web, require innovative information retrieval techniques to effectively satisfy users’ information need. A promising approach is to exploit document semantics in the IR process. For this purpose, high-quality semantic metadata is needed. This paper introduces a method to automatically create semantic metadata by using ontologically enhanced versions of common information extraction methods, such as named entity recognition and coreference resolution. Furthermore, this work also proposes the application of ontology-specific heuristic rules to further improve the quality of generated metadata. The results of our method was evaluated using a small test collection. High-level overview of the project (pdf file) - 400 KB as presented at the meeting for cultural heritage research projects in Luxembourg, on 30 June, 2006. ABSTRACT
Project Aims: - Bring digital cultural and scientific resources closer to their users.
- Create a flexible, reusable technology of visual contextualisation for images.
|  2008 Project Flyer
(pdf file) 1,1 MB  2008 ImageNotion Flyer
(pdf file) 1,5 MB 
2008 ImageNotion poster for Photokina fair (pdf file) 1,2 MB 
2008 project poster for Photokina fair (pdf file) 1,4 MB
2006 project Flyer (pdf file) 400 KB 
2006 project poster (gif file) 400 KB |
