Abstract. Active Appearance Models (AAM) are compact representations of the shape and appearance of objects. Fitting AAMs to images is a difficult, non-linear optimization task. Tr...
We present a novel method for inferring three-dimensional shape from a collection of defocused images. It is based on the observation that defocused images are the null-space of ce...
This paperpresentsa novel framework for streamingvideo over a Differentiated Services (DifPjew) network that jointly considers video source coding. packet classification and error...
Fan Zhai, Yiftach Eisenberg, Carlos E. Luna, Thras...
Many large scale physics-based simulations which take place on PC clusters or supercomputers produce huge amounts of data including vector fields. While these vector data such as ...
This paper presents a method for automatically converting raw GPS traces from everyday vehicles into a routable road network. The method begins by smoothing raw GPS traces using a...