THESEUS-Basistechnologien:
Evaluation
Organisationen:
- Fraunhofer-Institut für Digitale Medientechnologie (IDMT)
Kontakt:
Judith Liebetrau
Fraunhofer-Institut für Digitale Medientechnologie IDMT
Ehrenbergstraße 31
98693 Ilmenau
E-Mail: ltu@idmt.fhg.de
Zielsetzung:
Die Schwerpunkte der wissenschaftlich-technischen Arbeiten liegen bei der Evaluierung der von den Partnern im CTC entwickelten Technologien und Verfahren. Hierbei werden
u. a. Qualitäts-, Usability- und Sicherheitsaspekte untersucht. Die Aufgabe des CTC WP8 ist es vor allem, die in THESEUS entwickelten Technologien – z. B. Bilderkennung, Nutzerschnittstellen sowie die automatische Zuordnung von Metadaten – im Hinblick auf ihre Qualität zu beurteilen und zu bewerten. Dieser Evaluationsprozess generiert Rückmeldungen, die von den Entwicklern und Partnern in iterativer Anwendung im Design- und Entwicklungsprozess verwendet werden können. Zusätzlich wird die Evaluation einen Beitrag zur Verbesserung der Evaluationstechniken und zur Entwicklung von Standards liefern.
Präsentationen:
- Peter Dunker, Melanie Keller: ILLUMINATION NORMALIZATION FOR FACE RECOGNITION, A comparative study of conventional vs. perception-inspired algorithms. at BIOSIGNALS, Portugal, 2008
- Peter Dunker, Matthias Gruhne: Audio-Visual Fingerprinting: Components, Applications, Companies and an Example of Cross-Modal Media Identification, At ISCE, Portugal, 2008
- Peter Dunker, Juan José Bosch Vicente, Judith Liebetrau: Towards Evaluating the Core Technology Cluster of the German Research Project THESEUS, at QUAERO/ImageCLEF Workshop on Multimedia Information Retrieval Evaluation, Denmark, 2008 (Download)
- Stefanie Nowak, Peter Dunker: A generic evaluation framework for the evaluation of content-based image and video analysis tasks in THESEUS, at QUAERO/ImageCLEF Workshop on Multimedia Information Retrieval Evaluation, Denmark, 2008
- Fanny Klett and Helga Scholz: 360°View of Assessment and Quality Management, White
Paper SCORM 2.0 Workshop, USA, 2008 - Fanny Klett: Use of semantic and agent technologies towards personalization of information delivery, In S1000D harmonization meeting, Norway, 2008
- Stefanie Nowak: ImageCLEF LS-VCDT: Evaluation of multilabel image annotation incorporating domain knowledge and concept subjectivity, at CHORUS Final Conference Belgium, 2008
- Matthias Gruhne, Peter Dunker , Anton Mikhalev, Ilya, Fotis Andritsopoulos: Cross-modal identification of audiovisual streams directly from the compressed domain, at 13 International Symposium on Consumer Electronics ISCE’09, Japan, 2009
- Hanna Lukashevich, Stefanie Nowak, Peter Dunker: Using One-class SVM Outliers Detection for Verification of Collaboratively Tagged Image Training Sets, at International Conference on Multimedia and Expo (ICME 2009), Mexico, 2009
- Stefanie Nowak, Hanna Lukashevich: Multilabel Classification Evaluation using Ontology Information, At ESWC Workshop on Inductive Reasoning and Machine Learning on the Semantic Web, Greece, 2009
- Peter Dunker, Ronny Paduschek, Christian Dittmar, Stefanie Nowak and Matthias Gruhne: Evaluation of an Image and Music Indexing Prototype, at Workshop Audiovisuelle Medien, Germany, 2009
- Judith Liebetrau: The research program THESEUS and subjective quality assessment, at Current trends in Science Innovation, Greece, 2009
- Stefanie Nowak, Peter Dunker: A Consumer Photo Tagging Ontology:Concepts and Annotations, at THESEUS/ImageCLEF pre-Workshop 2009, Greece, 2009
- Stefanie Nowak, Peter Dunker: Overview of the CLEF 2009 Large-Scale Visual Concept Detection and Annotation Task, at CLEF Working Notes, Greece, 2009
- Alexander Loos: Evaluation of Algorithms for the Summarization of Photo Collections, at THESEUS/ImageCLEF pre-Workshop 2009, Greece, 2009 (Download)
Wissenschaftliche Publikationen:
2009
- Peter Dunker, Christian Dittmar, André Begau, Stefanie Nowak, Matthias Gruhne: Semantic High-Level Features for Automated Cross-Modal Slideshow Generation. 7th International Workshop on Content-Based Multimedia Indexing, CBMI, Chania, Greece, 2009. (Download)
- Peter Dunker, Ronny Paduschek, Christian Dittmar, Stefanie Nowak, Matthias Gruhne: Media Archive Management – Evaluation of an Image and Music Indexing Prototype. Workshop Audiovisuelle Medien, WAM, Chemnitz, Germany, 2009. (Download)
- Peter Dunker, Ronny Paduschek: Multimodal Fingerprinting for Repository Cleansing. 10th Workshop Digital Broadcasting, Ilmenau, Germany, 2009
- Matthias Gruhne, Peter Dunker, Anton Mikhalev, Ilya, Fotis Andritsopoulos: Cross-modal identification of audiovisual streams directly from the compressed domain. 13 International Symposium on Consumer Electronics, ISCE, Kyoto, Japan, 2009
- Matthias Gruhne, Peter Dunker, Ruben Tous: Using the MPEG Query Format for Cross-Modal Identification. Journal of Multimedia (JMM), Volume 4, Number 5, 2009 (Download)
- Alexander Loos, Ronny Paduschek, Daniel Kormann, Peter Dunker: Evaluation of Algorithms for the Summarization of Photo Collections. Theseus/ImageCLEF Preworkshop, Corfu, Greece, 2009. (Download)
- Hanna Lukashevich, Stefanie Nowak, Peter Dunker: Using One-class SVM Outliers Detection for Verification of Collaboratively Tagged Image Training Sets. IEEE International Conference on Multimedia and Expo, ICME, New York, USA, 2009. (Download)
- Stefanie Nowak, Hanna Lukashevich: Multilabel Classification Evaluation using Ontology Information. ESWC Workshop on Inductive Reasoning and Machine Learning on the Semantic Web, Heraklion, Greece, 2009. (Download)
- Stefanie Nowak, Peter Dunker: Overview of the CLEF 2009 Large-Scale Visual Concept Detection and Annotation Task, CLEF Working Notes 2009, Corfu, Greece, 2009 (Download)
- Stefanie Nowak, Peter Dunker: Visualization of Hierarchical Image Classification Evaluation. Theseus/ImageCLEF Preworkshop, Corfu, Greece, 2009
- Stefanie Nowak, Peter Dunker, Ronny Paduschek: THESEUS Meets ImageCLEF: Combining Evaluation Strategies for a New Visual Concept Detection Task 2009. CLEF 2008 Post-Workshop Proceedings. Evaluating Systems for Multilingual and Multimodal Information Access – 9th Workshop of the Cross-Language Evaluation Forum, CLEF 2008, Aarhus, Denmark, September 17-19, 2008, Revised Selected Papers. LNCS, Available soon. (Download)
2008
- Peter Dunker, Melanie Keller: Illumination Normalization for Face Recognition, a Comparative Study of Conventional vs. Perception-Inspired Algorithms. International Conference on Bio-inspired Systems and Signal Processing, Biosignals, Funchal, Portugal, 2008. (Download)
- Peter Dunker, Melanie Keller: Diffusion Filtering for Illumination Invariant Face Recognition, Illumination Approximation with Diffusion Filters within Retinex Context. 3rd International Conference on Computer Vision Theory and Applications, VISAPP, Funchal, Portugal, 2008
- Peter Dunker, Matthias Gruhne: Audio-Visual Fingerprinting: Components, Applications, Companies and an Example of Cross-Modal Media Identification. 12th Annual IEEE International Symposium on Consumer Electronics, ISCE, Algarve, Portugal, 2008
- Peter Dunker, Matthias Gruhne: Personal Television: A Crossmodal Analysis Approach. 12th Annual IEEE International Symposium on Consumer Electronics, ISCE, Algarve, Portugal, 2008. (Download)
- Peter Dunker, Stefanie Nowak, André Begau, Cornelia Lanz: Content-based Mood Classification for Photos and Music: A generic multi-modal classification framework and evaluation approach. International Conference on Multimedia Information Retrieval, MIR, Vancouver, Canada, 2008. (Download)
- Peter Dunker, Juan José Bosch Vicente, Judith Liebetrau: Towards Evaluating the Core Technology Cluster of the German Research Project THESEUS. QUAERO/ImageCLEF Workshop on Multimedia Information Retrieval Evaluation, Aarhus, Denmark, 2008 (Abstract), (Präsentation)
- Fanny Klett: The Rising Challenge for Assessment Standardization in Web Based Learning, Education and Training. 16th International Conference on Computers in Education, ICCE, Taipei, Taiwan, 2008 (Download)
- Fanny Klett and Helga Scholz: 360° View of Assessment and Quality Management, White Paper. SCORM 2.0 Workshop, Pensacola, USA, 2008 (Download)
- Fanny Klett: Use of semantic and agent technologies towards personalization of information delivery. S1000D harmonization meeting, Norway, 2008.
- Stefanie Nowak, Peter Dunker: A generic evaluation framework for the evaluation of content-based image and video analysis tasks in THESEUS. QUAERO/ImageCLEF Workshop on Multimedia Information Retrieval Evaluation, Aarhus, Denmark, 2008. (Download)
Software, Softwarebeschreibungen und Dokumentationen:
- Erzeugung Video- und Bild-Distortions
Das Fraunhofer IDMT hat ein Tool entwickelt, mit dem man vielfältige Video-Distortions generieren kann. Mit dem „Video Manipulator“ können beispielsweise neben der Störung durch verschiedenartiges Rauschen die Videos geometrisch verändert (Cropping, Verzerrung) und die Schärfe beeinflusst werden. - THESEUS-Evaluationsframework Face Detector
- THESEUS-Evaluationsframework Person Detection
- THESEUS-Evaluationsframework Bild-Segmentierung
- Photo-Annotationstool
Es wurde ein Annotationstool in C++ und ein Evaluationsmaß in Java programmiert. Das Annotationstool erlaubt, die Fotos mit den 53 Konzepten zu annotieren und wurde für die Groundtruth-Generierung genutzt. - Picture und Quality Assessment GUI
GUI zur Durchführung und automatischen Ergebnisspeicherung von visuellen Tests. Die GUI unterstützt aktuell nur zwei verschiedene Methoden, wird im Laufe des THESEUS-Projekts stetig erweitert.
Wichtige Hyperlinks:
- THESEUS auf der Webseite idmt.fhg.de
- THESEUS auf der Webseite imageclef.org
- Webseite des THESEUS-Programms