In this report, we propose a statistical model to deal with the discrete-distribution data varying over time. The proposed model – HMM+DM – extends the Dirichlet mixture model to the dynamic case: Hidden Markov Model with Dirichlet mixture output. Both the inference and parameter estimation procedures are proposed. Experiments on the generated data verify the proposed algorithms. Finally, we di...

We describe the maximum-likelihood parameter estimation problem and how the ExpectationMaximization (EM) algorithm can be used for its solution. We first describe the abstract form of the EM algorithm as it is often given in the literature. We then develop the EM parameter estimation procedure for two applications: 1) finding the parameters of a mixture of Gaussian densities, and 2) finding the...

In this paper we present an extension of our previously proposed feature space stereo-based stochastic mapping (SSM). As distinct from an auxiliary stereo Gaussian mixture model in the front-end in our previous work, a stereo HMM model in the back-end is used. The basic idea, as in feature space SSM, is to form a joint space of the clean and noisy features, but to train a Gaussian mixture HMM i...

In this paper we describe a framework of analyzing programs belonging to different TV program genres using Hidden Markov Models and pseudo-semantic features derived from video shots. Clustering using Gaussian mixture models is used to determine the order of the models. Results for initial genre classification experiments using two simple features derived from video shots are given.

The recognition of the emotional states of speaker is a multidisciplinary research area that has received great interest in the last years. One of the most important goals is to improve the voiced-based human-machine interactions. Recent works on this domain use the prosodic features and the spectrum characteristics of speech signal, with standard classifier methods. Furthermore, for traditiona...

Two approaches are proposed for the design of tied-mixture hidden Markov models (TMHMM). One approach improves parameter sharing via partial tying of TMHMM states. To facilitate tying at the substate level, the state emission probabilities are constructed in two stages or, equivalently, are viewed as a “mixture of mixtures of Gaussians.” This paradigm allows, and is complemented with, an optimi...

We present a system to beautify curves: i.e. to take curves that roughly depict some property of interest and make them look more like what experts would draw. The focus of our work is in applications for artistic drawings, but our system can also be plugged into various other domains where curves and trajectories play a dominant role such as robot path planning, animation or edge-deblurring. O...

This article reports a two-part study of structured acoustic modeling of speech. First, speaker-independent clustering of speech material was used as the basis for a practical cluster-based acoustic modeling. Each cluster’s training material is applied to the adaptation of baseline hidden Markov model (HMM) parameters for recognition purposes. Further, the training material of each cluster is a...

A linear latent growth curve mixture model is presented which includes switching between growth curves. Switching is accommodated by means of a Markov transition model. The model is formulated with switching as a highly constrained multivariate mixture model and is fitted using the freely available Mx program. The model is illustrated by analyzing data from the National Longitudinal Survey of Y...

Text-to-speech has long been centered on the production of an intelligible message of good quality. More recently, interest has shifted to the generation of more natural and expressive speech. A major issue of existing approaches is that they usually rely on a manual annotation in expressive styles, which tends to be rather subjective. A typical related issue is that the annotation is strongly ...