MIR 2005 Special Sessions
Session 1: Machine Learning for Visual Information Retrieval
Chair: Nozha Boujemaa, INRIA, France
This special session is sponsored by MUSCLE Network of Excellence
In the recent years, machine learning based methods appears as a key
reliable foundation to provide efficient solution for numerous problems in
visual information description and retrieval by content. This special
session provides a representative tour of different machine learning methods
contributions in video and still images collections: such as mental face
retrieval through an information theory framework providing a feedback
mechanism for target search, cluster-wise semantics endorsement approach to
enhance the retrieval accuracy, semi-supervised active clustering of image
signature space, statistical similarity measure that provides efficient
sub-linear search in copy detection problem and finally a Bayesian inference
based-method for event detection in videos.
Presentations in the session:
- A Cluster-wise Semantics Endorsement Approach to Image Retrieval
Jia Li
The Pennsylvania State University, University Park, PA, USA
- Exploiting Temporal Discontinuities for Event Detection and Manipulation
in Video Streams
Hugh Denman, Erika Doyle, Anil Kokaram, Daire Lennon, Rozenn Dahyot, Ray Fuller
University of Dublin, Trinity College, Dublin 2, Ireland
- An Interactive System for Mental Face Retrieval
Yuchun Fang (1,2), Donald Geman (3,1), Nozha Boujemaa (1)
(1) INRIA, France, (2) Shanghai University, China, and (3) Johns Hopkins University, USA
- Discriminant Local Features Selection using Efficient
Density Estimation in a Large Database
Alexis Joly, Olivier Buisson
INRIA, France and INA - National Audiovisual Institute, France
- Hybrid Visual and Conceptual Image Representation within Active
Relevance Feedback Context
Marin Ferecatu, Nozha Boujemaa, Michel Crucianu
IMEDIA Group - INRIA Rocquencourt, France
Session 2: Multimedia Information Retrieval: Challenges and Real-world Applications
Chair: James Z. Wang, The Pennsylvania State University, USA
Researchers in multimedia, databases, computer vision, machine
learning, signal and image processing, and statistics have worked on
multimedia information retrieval for over a decade. A number of
significant technological advances have been achieved in this field,
including region-based image retrieval, automatic linguistic indexing
of pictures by concept modeling, and feature space indexing. Some
of the techniques have been applied to real-world application areas
such as art image retrieval, biomedical image and video retrieval,
sensor networks, large-scale online photo sharing communities,
classification of objectionable images on the Web, industrial
processes, and computer forensics. We envision more and more
real-world applications of multimedia information retrieval to appear
in the coming years. In this session, we will invite active
researchers in this area to showcase latest technology
advances and to discuss future possibilities.
Presentations in the session:
- Evaluation Strategies for Image Understanding and Retrieval
Keiji Yanai, Nikhil Shirahatti, Prasad Gabbur, Kobus Barnard
University of Arizona, USA
- Localized Content-Based Image Retrieval
Rouhollah Rahmani, Sally Goldman, Hui Zhang, Jason Fritts
Washington University at St. Louis, USA
- A Content-Based Image Retrieval System for Fish Taxonomy
Yixin Chen, Henry L. Bart, Jr., Fei Teng
University of New Orleans and Tulane University Museum of Natural History, USA
- Extracting Information from Multimedia Meeting Collections
Daniel Gatica-Perez, Dong Zhang, and Samy Bengio
IDIAP Research Institute, Switzerland
- Content-Based Image Retrieval - Approaches and Trends in the New Age
Ritendra Datta, Jia Li, James Z. Wang
The Pennsylvania State University, University Park, PA, USA
Note:
For questions, ask the
Chair for Special Sessions, James Z. Wang, jwang at ist.psu. edu .