PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Multimodal Information Fusion
Norman Poh, Thirimahos Bourlai and Josef Kittler
In: Multimodal Signal Processing (2010) Academic Press , pp. 153-169. ISBN 978-0-12-374825-6


This chapter introduces the concept of "adaptive information fusion" -- an approach that combines the information from several classifiers that takes into account the signal quality derived from a sample. This approach is both intuitive and appealing. This is because a particular source of noise may only corrupt the information derived from a particular HCI modality. For instance, while the the face modality can be affected by the changing lighting conditions, the speech modality is affected only by the audible environmental noise sources (e.g., car passing by). This chapter introduces different levels of fusion, adaptive fusion strategies and outlines future research issues in this direction

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Book Section
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Multimodal Integration
ID Code:6857
Deposited By:Norman Poh
Deposited On:08 April 2010