PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Partial relevance in interactive facial image retrieval
Zhirong Yang and Jorma Laaksonen
In: 3rd International Conference on Advances in Pattern Recognition, 22-25 Aug 2005, Bath, United Kingdom.

Abstract

For databases of facial images, where each subject has only a few images, the query precision of interactive retrieval suffers from the problem of extremely small class sizes. A novel method is proposed to relieve this problem by applying partial relevance to the interactive retrieval. This work extends an existing content-based image retrieval system, PicSOM, by relaxing the relevance criterion in the early rounds of the retrieval. Moreover, we apply linear discriminant analysis as a preprocessing step before training the Self-Organizing Maps (SOMs) so that the resulting SOMs have stronger discriminative power. The results of simulated retrieval experiments suggest that for semantic classes such as ``black persons'' or "bearded persons" the first image which depicts the target subject can be obtained three to six times faster than by retrieval without the partial relevance.

EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Machine Vision
User Modelling for Computer Human Interaction
Information Retrieval & Textual Information Access
ID Code:1730
Deposited By:Jorma Laaksonen
Deposited On:28 November 2005