Dynamic face pose change and noise make it difficult to track multi-view faces in a cluttering environment. In this paper, we propose a graphical model based method, which combin...
This paper presents a Bayesian network based multimodal fusion method for robust and real-time face tracking. The Bayesian network integrates a prior of second order system dynami...
We present the IBM systems submitted and evaluated within the CLEAR'06 evaluation campaign for the tasks of single person visual 3D tracking (localization) and 2D face trackin...
: This paper describes a new approach for the creation of an adaptive system able to selectively combine dynamic multidimensional information sources to perform state estimation. T...
In developing automated systems to recognize the emotional content of music, we are faced with a problem spanning two disparate domains: the space of human emotions and the acoust...
Erik M. Schmidt, Douglas Turnbull, Youngmoo E. Kim