This paper addresses the problem of tracking objects which undergo rapid and significant appearance changes. We propose a novel coupled-layer visual model that combines the targe...
This paper presents a novel sequence labeling model based on the latent-variable semiMarkov conditional random fields for jointly extracting argument roles of events from texts. ...
Computational and comparative analysis of protein-protein interaction (PPI) networks enable understanding of the modular organization of the cell through identification of functio...
Image-based and model-based methods are two representative rendering methods for generating virtual images of objects from their real images. Extensive research on these two metho...
We present a model-based method for accurate extraction of pedestrian silhouettes from video sequences. Our approach is based on two assumptions, 1) there is a common appearance t...