Dynamic face pose change and noise make it difficult to track multi-view faces in a cluttering environment. In this paper, we propose a graphical model based method, which combin...
Technical support procedures are typically very complex. Users often have trouble following printed instructions describing how to perform these procedures, and these instructions...
Tessa A. Lau, Lawrence D. Bergman, Vittorio Castel...
In this paper, a motion-based approach for detecting highlevel semantic events in video sequences is presented. Its main characteristic is its generic nature, i.e. it can be direc...
Background: Traditional algorithms for hidden Markov model decoding seek to maximize either the probability of a state path or the number of positions of a sequence assigned to th...
We propose and analyze a distribution learning algorithm for variable memory length Markov processes. These processes can be described by a subclass of probabilistic nite automata...