Interactive Hierarchical SOM for Image Retrieval Visualization

Authors:
Yi Liu;Masahiro Takatsuka
Affiliations:
ViSLAB, The School of IT, The University of Sydney, Australia 2006;ViSLAB, The School of IT, The University of Sydney, Australia 2006
Venue:
ICONIP '09 Proceedings of the 16th International Conference on Neural Information Processing: Part II
Year:
2009

Citing 3
Cited 0

A multilayer self-organizing feature map for range image segmentation

Neural Networks
Very Large Two-Level SOM for the Browsing of Newsgroups

ICANN 96 Proceedings of the 1996 International Conference on Artificial Neural Networks
Spherical self-organizing map using efficient indexed geodesic data structure

Neural Networks - 2006 Special issue: Advances in self-organizing maps--WSOM'05

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents an interactive hierarchical visualization system for an image retrieval application. This visualization system needs to present the similarities of images. Furthermore, it is required to provide an easy way to explore and navigate images' feature space at different levels of detail. Our system utilizes a Multi-layer Geodesic Self-Organizing Map (GeodesicSOM) and Learning Vector Quantization (LVQ) to increase the accuracy in data representation at different levels of detail. The Multi-layer GeodesicSOM provides fast access/navigation to a large amount of image data while the LVQ rectifies the inconsistency in topological data representation between different layers.