The dynamic texture (DT) is a probabilistic generative model, defined over space and time, that represents a video as the output of a linear dynamical system (LDS). The DT model ...
In this paper, we propose an optimization framework for real-time voice transmission over wireless tandem channels prone to both bit errors and packet erasures. Utilizing a hybrid...
—ModelNet is a network emulator designed for repeatable, large-scale experimentation with real networked systems. This talk introduces the ideas behind ModelNet that have made it...
Kashi Venkatesh Vishwanath, Amin Vahdat, Ken Yocum...
For many computer vision problems, it is very important to produce the groundtruth data. Manual data labeling is labor-intensive and prone to the human errors, whereas fully autom...
Most existing methods of stereo matching focus on dealing with clear image pairs. Consequently, there is a lack of approaches capable of handling degraded images captured under ch...