— Wireless local area networks offer a range of transmitted data rates that are to be selected according to estimated channel conditions. However, due to packet overheads and con...
A master/worker paradigm for executing large-scale parallel discrete event simulation programs over networkenabled computational resources is proposed and evaluated. In contrast t...
We present an emotion recognition system based on a probabilistic approach to adaptive processing of Facial Emotion Tree Structures (FETS). FETS are made up of localized Gabor fea...
The extraction of optimal features, in a classification sense, is still quite challenging in the context of large-scale classification problems (such as visual recognition), inv...
Abstract. Tuning hyper-parameters is a necessary step to improve learning algorithm performances. For Support Vector Machine classifiers, adjusting kernel parameters increases dra...