This paper presents an empirical study of the joint wavelet statistics for textures and other random imagery. There is a growing realization that modeling wavelet coefficients as independent, or at best correlated only across scales, assuming independence within a scale, may be a poor assumption. While recent developments in wavelet-domain Hidden Markov Models (notably HMT-3S) account for withi...

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, VOL. XX, NO. Y, MONTH 1999 On the Complexity of Explicit Duration HMMs Carl Mitchell, Mary Harper, and Leah Jamieson Abstract|We introduce a new recursion that reduces the complexity of training a semi-Markov model with continuous output distributions. We show that the cost of training is proportional to M2 +D, compared to M2D with the standard ...

A score-following algorithm for polyphonic MIDI performances is presented that can handle performance mistakes, ornaments, desynchronized voices, arbitrary repeats and skips. The algorithm is derived from a stochastic performance model based on hidden Markov model (HMM), and we review the recent development of model construction. In this paper, the model is further extended to capture the multi...

We propose the use of Markov Logic Networks (MLNs) as a highly flexible and expressive formalism for the harmonic analysis of audio signals. Using MLNs information about the physical and semantic content of the signal can be intuitively and compactly encoded and expert knowledge can be easily expressed and combined using a single unified formal model that combines probabilities and logic. In pa...

This paper gives a method for computing distributions associated with patterns in the state sequence of a hidden Markov model, conditional on observing all or part of the observation sequence. Probabilities are computed for very general classes of patterns (competing patterns and generalised later patterns), and thus the theory includes as special cases results for a large class of problems tha...

In this paper we present a novel approach to generate a sequence of head motion units given some speech. The modelling approach is based on the notion that head motion can be divided into a number of short homogeneous units that can be modelled individually. The system is based on Hidden Markov Models (HMM), which are trained on motion units and act as a sequence generator. They can be evaluate...

Existing works on user behavior analysis mainly focus on modeling a single behavior and predicting whether a user will take an action or not. However, users’ behaviors do not always happen in isolation, sometimes, different behaviors may happen simultaneously. Therefore, in this paper, we try to analyze the combination of basic behaviors, called behavioral state here, which can describes users’...

This paper highlights the estimation and test procedures for multi-state Markov models with covariate dependences in higher orders. Logistic link functions are used to analyze the transition probabilities of three or more states of a Markov model emerging from a longitudinal study. For illustration purpose the models are used for analysis of panel data on Health and Retirement Study conducted i...

We describe a simple transformation that allows for the fast recovery of a tree from the probabilities such a tree induces on the colourations of its leaves under a simple Markov process (with unknown parameters). This generalizes earlier results by not requiring the transition matrices associated with the edges of the tree to be of a particular form, or to be related by some fixed rate matrix,...