A new probabilistic background model based on a Hidden Markov Model is presented. The hidden states of the model enable discrimination between foreground, background and shadow. Th...
In modern automatic speech recognition systems, it is standard practice to cluster several logical hidden Markov model states into one physical, clustered state. Typically, the cl...
Our research focuses on analysing human activities according to a known behaviorist scenario, in case of noisy and high dimensional collected data. The data come from the monitori...
Visual vocabulary serves as a fundamental component in many computer vision tasks, such as object recognition, visual search, and scene modeling. While state-of-the-art approaches...
In this paper, we present an Arabic morphological analysis system that assigns, for each word of an unvoweled Arabic sentence, a unique root depending on the context. The proposed...