AIM@SHAPE KSPACE CNR ICT
Classification and Annotation of Multidimensional Content
Motivation and Objectives

Interpreting and classifying multidimensional media is a crucial aspect for content understanding and retrieval as well as for application specific analysis. Here multidimensionality is understood not only as the intrinsic dimensionality of a media content (e.g. 2D images or 3D objects) but also refers to the combination of a variety of heterogeneous signals concurring to describe a single phenomenon, (e.g. fusion of multispectral and panchromatic images). The significant scientific and technological advances in media content acquisition and generation make available a huge amount of assorted and multidimensional data which is becoming more and more accessible also to non expert users. This richness of data increases the amount of information, but also calls for the development of suitable tools for their combination and understanding to achieve full data understanding and exploitation.

The project Multimodal and Multidimensional Content and Media of the Italian National Research Council deals with fundamental aspects associated to media creation, processing, classification, related knowledge formalization and annotation that are necessary for properly facing the above challenges.

This special session aims at reporting part of the significant results of this project by presenting some basic issues concerning media classification, management and retrieval: from an algorithmic point of view, e.g. data fusion and processing, to application-driven data interpretation, to multimedia semantics and ontology formalization and specification. Some of the contributions include also results of research carried out in the frame of other EU initiatives, as the AIM@SHAPE and MUSCLE Networks of Excellence.

Topics

Topics of interest include, but are not limited to:

  • Context-dependent classification of images
  • Data fusion
  • 3D object classification
  • Multimedia Semantics and Ontologies
  • Ontology and standards
Session Outline

This special session includes contributions dealing with some fundamental aspects of media analysis, classification, management and retrieval: from algorithms and computational aspects to semantics and ontology specification.

Some papers investigate on multidimensional content classification and data integration in the remote sensing context. Other contributions face the problem of medical image fusion, analysis and classification. In this frame, proposals for 3D modelling and content classification based on dissimilarity measures among parts are presented. Furthermore, some papers discuss aspects concerning media metadata management and the exploitation of ontology for resources and content semantic annotation.

Contact information for the Special Session chairs

Franca Giannini
Institute for Applied Mathematics and Information Technologies
Consiglio Nazionale delle Ricerche
Via de Marini, 6
16149 Genova - Italy
Franca.giannini@ge.imati.cnr.it

Ovidio Salvetti
Institute of Information Science and Technologies
Consiglio Nazionale delle Ricerche
Via G. Moruzzi,
56124 Pisa - Italy
Ovidio.Salvetti@isti.cnr.it

Accepted Papers
  • Semantic Annotation of 3D Surface Meshes based on Feature Characterization
    M. Attene, F. Robbiano, M. Spagnuolo, B. Falcidieno
  • A Wavelet-based Algorithm for Multimodal Medical Image Fusion (short paper)
    B. Alfano, M. Ciampi, G. De Pietro
  • Thesaurus-based Ontology on Image Analysis (short paper)
    S. Colantonio, I. Gurevich, M. Martinelli, O. Salvetti, Y. Trusova
  • Context-Sensitive Pan-Sharpening of Multispectral Images (short paper)
    B. Aiazzi, L. Alparone, S. Baronti, A. Garzelli, F. Lotti, F. Nencini, M. Selva
  • 3D Classification via Structural Prototypes (short paper)
    S. Biasotti, D. Giorgi, S. Marini, M. Spagnuolo, B. Falcidieno