A neuronet approach to information fusion
Thomas S. Huang, C.P. Hess, Hao Pan, Zhi‐Pei Liang · 2002
Neuronet approaches offer a unique and powerful tool for nonlinear information fusion. Unlike traditional techniques, neuronets do not require explicit environmental models or descriptions of sensor characteristics. This paper describes a technique for sensor fusion which makes use of a new neural model to combine data autonomously extracted from different sources. Application of the technique to bimodal recognition of combined speech/image signals is discussed.