We derive an efficient learning algorithm for model-based source separation for use on single channel speech mixtures where the precise source characteristics are not known a pri...
Segmenting meshes into natural regions is useful for model understanding
and many practical applications. In this paper, we present
a novel, automatic algorithm for segmenting me...
Yu-Kun Lai, Qian-Yi Zhou, Shi-Min Hu, Ralph R. Mar...
Abstract. Navigation for and tracking of humans within a building usually implies significant infrastructure investment and devices are usually too high in weight and volume to be...
An arti"cial neural network (ANN) algorithm is proposed that incorporates both market segmentation and discriminant (regression) analysis of the segments. The method simultan...
A neural network approach is presented for modeling and characterization of on-chip copper spiral inductors. The approach involves the creation of neural network models to map 3D ...