In this paper we describe the major elements of MIT Lincoln Laboratory’s Gaussian mixture model (GMM)-based speaker verification system used successfully in several NIST Speaker Recognition Evaluations (SREs). The system is built around the likelihood ratio test for verification, using simple but effective GMMs for likelihood functions, a universal background model (UBM) for alternative speaker...

In this work a sound transformation model based on Gaussian Mixture Models is introduced and evaluated for audio morphing. To this aim, the GMM is used to build the acoustic model of the source sound, and a set of conversion functions, which rely on the acoustic model, is used to transform the source sound. The method is experimented on a set of monophonic sounds and results show that it provid...

In transform image coding, the histograms of transform coeecients can be approximately modeled by generalized Gaussian (GG) random variables. However, the GG models may not t the DC distribution. One approach uses DPCM for the DC data, which greatly complicates bit allocation; another assumes a single Gaussian (SG) model, which may be a poor model. As an alternative, this paper proposes a nite ...

Joint uncertainty decoding (JUD) is an effective model-based noise compensation technique for conventional Gaussian mixture model (GMM) based speech recognition systems. In this paper, we apply JUD to subspace Gaussian mixture model (SGMM) based acoustic models. The total number of Gaussians in the SGMM acoustic model is usually much larger than for conventional GMMs, which limits the applicati...

Mixtures of Gaussians are a crucial statistical modeling tool at the heart of many challenging applications in computer vision and machine learning. In this paper, we first describe a novel and efficient algorithm for simplifying Gaussian mixture models using a generalization of the celebrated k-means quantization algorithm tailored to relative entropy. Our method is shown to compare experiment...

This paper presents a new algorithm for unsupervised incremental learning based on a Bayesian framework. The algorithm, called IGMM (for Incremental Gaussian Mixture Model), creates and continually adjusts a Gaussian Mixture Model consistent to all sequentially presented data. IGMM is particularly useful for on-line incremental clustering of data streams, as encountered in the domain of mobile ...

This paper discusses a method for performing independent component analysis exploiting Gaussian mixture models (GMMs). Previously most techniques that combine these methods have used GMMs to model the source signals. This paper considers a parsimonious method for modelling the observed signals. The GMM is fitted to the observed data using a modified version of the expectation maximisation algor...

The current paper proposes a novel color correction approach based on a probabilistic segmentation framework by using 3D Gaussian Mixture Models. Regions are used to compute local color correction functions, which are then combined to obtain the final corrected image. The proposed approach is evaluated using both a recently published metric and two large data sets composed of seventy images. Th...