In this paper, we propose a novel predictive model for
object boundary, which can integrate information from any
sources. The model is a dynamic “object” model whose
manifes...
Tian Shen (Lehigh University), Hongsheng Li (Lehig...
Abstract. We propose an efficient way to account for spatial smoothness in foreground-background segmentation of video sequences. Most statistical background modeling techniques re...
Pedestrian navigation is a complex function of human dynamics, a desired destination, and the presence of obstacles. People cannot stop and start instantaneously and their turning...
Current source based cell models are becoming a necessity for accurate timing and noise analysis at 65nm and below. Voltage waveform shapes are increasingly more difficult to repr...
Many low-level vision algorithms assume a prior probability over images, and there has been great interest in trying to learn this prior from examples. Since images are very non G...