A self-organizing neural model for multimedia information fusion
Luong-Dong Nguyen, Kia Yan. Woon, Ah‐Hwee Tan · Singapore Management University Institutional Knowledge (InK) (Singapore Management University) · 2008
This paper presents a self-organizing network model for the fusion of multimedia information. By synchronizing the encoding of information across multiple media channels, the neural model known as fusion adaptive resonance theory (fusion ART) generates clusters that encode the associative mappings across multimedia information in a real-time and continuous manner. In addition, by incorporating a semantic category channel, fusion ART further enables multimedia information to be fused into predefined themes or semantic categories. We illustrate the fusion ARTpsilas functionalities through experiments on two multimedia data sets in the terrorist domain and show the viability of the proposed approach.