Abstract--We investigate an abstraction method, called meanfield method, for the performance evaluation of dynamic networks with pairwise communication between nodes. It allows us ...
Many computer vision problems can be formulated as low rank bilinear minimization problems. One reason for the success of these problems is that they can be efficiently solved usin...
Decomposition theorems in classical Fourier analysis enable us to express a bounded function in terms of few linear phases with large Fourier coefficients plus a part that is pseu...
Automated analysis of human affective behavior has attracted increasing attention in recent years. With the research shift toward spontaneous behavior, many challenges have come to...
This paper presents a wavelet neural-network for learning and approximation of chaotic time series. Wavelet-networks are inspired by both feed-forward neural networks and the theo...