Doctoral thesis
OA Policy
English

Information fusion for multimedia: exploiting feature interactions for semantic feature selection and construction

ContributorsKludas, Jana
Defense date2010-12-10
Abstract

The joint processing of multimedia data received a lot of attention by computer science engineers and researchers in the last decade. This is for once due to the flood of multimedia data such as videos and photographs that are easily created nowadays due to ubiquitous capturing devices like cameras in mobile phones and that can be instantly published in the Internet. Secondly, the success of content-only processing approaches that only use the visual modality stayed behind expectations so far. Therefore, the fusion of multiple information sources, so called modalities, came into the focus of interest. Other application areas of multi modal fusion are, for example, biometric identification and emotion recognition systems, where it is hoped to increase accuracy and reliability over mono modal systems.

Keywords
  • Multimodal Information Fusion
  • Attribute Relevance Detection
  • Multivariate Information Measures
  • Semantic Learning
  • Attribute Selection and Construction
Citation (ISO format)
KLUDAS, Jana. Information fusion for multimedia: exploiting feature interactions for semantic feature selection and construction. Doctoral Thesis, 2010. doi: 10.13097/archive-ouverte/unige:14539
Main files (1)
Thesis
accessLevelPublic
Identifiers
913views
812downloads

Technical informations

Creation15/02/2011 14:50:00
First validation15/02/2011 14:50:00
Update time14/03/2023 17:13:13
Status update14/03/2023 17:13:13
Last indexation29/10/2024 18:52:06
All rights reserved by Archive ouverte UNIGE and the University of GenevaunigeBlack