J. Robert Boston

Learn More
Hidden Markov modeling (HMM) provides a probabilistic framework for modeling a time series of multivariate observations. An HMM describes the dynamic behavior of the observations in terms of movement among the states of a nite-state machine. In this paper, we present an algorithm that selects an HMM topology for a set of time series data. Our method selects(More)
Temporomandibular joint (TMJ) sounds and motion were recorded during two clinically-derived movements--simple jaw opening and jaw protrusion followed by opening--from ten patients. A new time-frequency method--radially Gaussian kernel distribution--was applied to classify the TMJ clicking sounds into six groups, type I to type VI, based on the(More)