Clay Spence - Academia.edu (original) (raw)

Uploads

Papers by Clay Spence

Research paper thumbnail of Analyzing hyperspectral images into multiple subspaces using Gaussian mixture models

Research paper thumbnail of Hirarchical image probability (HIP) model

Research paper thumbnail of Recipes for the linear analysis of EEG

Research paper thumbnail of Response error correction-a demonstration of improved human-machine performance using real-time EEG monitoring

IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2003

Research paper thumbnail of Supervised learning of detection and classification tasks with uncertain training data

Research paper thumbnail of a Thomas-Fermi Approximation for Fermions in Lattice Gauge Theory

The inclusion of fermions in lattice gauge theory presents us with the problems of species doubli... more The inclusion of fermions in lattice gauge theory presents us with the problems of species doubling and explicit chiral symmetry breaking. In this dissertation Thomas -Fermi approximations are derived for the effective action due to fermions and the chiral order parameter in the continuum. They are then written in terms of lattice gauge field variables, and used in Monte Carlo

Research paper thumbnail of Thomas-Fermi approximation to functional determinants in external gauge potentials

Research paper thumbnail of Method and system for filtering, registering, and matching 2.5D normal maps

Research paper thumbnail of Convolutive blind source separation using a multiple decorrelation method

Research paper thumbnail of Method and apparatus for computing the probability distribution of an image having local and non-local information

Research paper thumbnail of On-line Blind Source Separation of Non-Stationary Signals

We have shown previously that non-stationarysignals recorded in a static multi-path environment c... more We have shown previously that non-stationarysignals recorded in a static multi-path environment can oftenbe recovered by simultaneously decorrelating varyingsecond order statistics. As typical sources are often moving,however, the multi-path channel is not static. We presenthere an on-line gradient algorithm with adaptive step sizein the frequency domain based on second derivatives, whichwe refer to as multiple adaptive decorrelation (MAD). Wecompared the

Research paper thumbnail of Geometric source preparation signal processing technique

Research paper thumbnail of Method and apparatus for image processing by generating probability distribution of images

Research paper thumbnail of Separation of Non-Stationary Natural Signals

CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of sourc... more CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of source separation use the assumption of statistical independence. To capture statistical independence higher order statistics are required. In this chapter we will demonstrate how higher ...

Research paper thumbnail of Convolutive Source Separation and Signal Modeling with ML

In independent component analysis (ICA) an instantaneous mix of sources can be recovered using ma... more In independent component analysis (ICA) an instantaneous mix of sources can be recovered using maximum Likelihood (ML). In Convolutive blind source separation (BSS) the mixture arises as a combination of differently convolved source signals due to time delays and a reverberating acoustic environment. Instead of modeling a particular time instant now a time window of the mixed signals has to be modeled. This allows to combine ICA with traditional ML signal modeling techniques. Here we use an auto-regressive (AR) model of the sources leading to a generalization of contextual ICA [18] to the convolutive case. This may improve source separation avoiding the typical whitening of the sources, and may allow us to incorporation simultaneous enhancing of the signal based on the AR models. 1 Introduction Independent component analysis (ICA) aims to find statistical independent signals in an instantaneous linear mix. This concept was first introduced and formalized by Comon [7]. In recent year...

Research paper thumbnail of On-line Blind Source Separation of Non-Stationary Signals

We have shown previously that non-stationarysignals recorded in a static multi-path environment c... more We have shown previously that non-stationarysignals recorded in a static multi-path environment can oftenbe recovered by simultaneously decorrelating varyingsecond order statistics. As typical sources are often moving,however, the multi-path channel is not static. We presenthere an on-line gradient algorithm with adaptive step sizein the frequency domain based on second derivatives, whichwe refer to as multiple adaptive decorrelation (MAD). Wecompared the

Research paper thumbnail of An Iterative Method for Vector Median Filtering

2007 IEEE International Conference on Image Processing, 2007

Research paper thumbnail of Apparatus for separating convolutively mixed signals derived from concurrent microphone recordings into plurality of sources using geometric information and adaptive beam forming

Research paper thumbnail of Separation of non-stationary natural signals

Principles and Practice, 2001

CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of sourc... more CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of source separation use the assumption of statistical independence. To capture statistical independence higher order statistics are required. In this chapter we will demonstrate how higher ...

Research paper thumbnail of Convolutive blind separation of non-stationary sources

IEEE Transactions on Speech and Audio Processing, 2000

Research paper thumbnail of Analyzing hyperspectral images into multiple subspaces using Gaussian mixture models

Research paper thumbnail of Hirarchical image probability (HIP) model

Research paper thumbnail of Recipes for the linear analysis of EEG

Research paper thumbnail of Response error correction-a demonstration of improved human-machine performance using real-time EEG monitoring

IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2003

Research paper thumbnail of Supervised learning of detection and classification tasks with uncertain training data

Research paper thumbnail of a Thomas-Fermi Approximation for Fermions in Lattice Gauge Theory

The inclusion of fermions in lattice gauge theory presents us with the problems of species doubli... more The inclusion of fermions in lattice gauge theory presents us with the problems of species doubling and explicit chiral symmetry breaking. In this dissertation Thomas -Fermi approximations are derived for the effective action due to fermions and the chiral order parameter in the continuum. They are then written in terms of lattice gauge field variables, and used in Monte Carlo

Research paper thumbnail of Thomas-Fermi approximation to functional determinants in external gauge potentials

Research paper thumbnail of Method and system for filtering, registering, and matching 2.5D normal maps

Research paper thumbnail of Convolutive blind source separation using a multiple decorrelation method

Research paper thumbnail of Method and apparatus for computing the probability distribution of an image having local and non-local information

Research paper thumbnail of On-line Blind Source Separation of Non-Stationary Signals

We have shown previously that non-stationarysignals recorded in a static multi-path environment c... more We have shown previously that non-stationarysignals recorded in a static multi-path environment can oftenbe recovered by simultaneously decorrelating varyingsecond order statistics. As typical sources are often moving,however, the multi-path channel is not static. We presenthere an on-line gradient algorithm with adaptive step sizein the frequency domain based on second derivatives, whichwe refer to as multiple adaptive decorrelation (MAD). Wecompared the

Research paper thumbnail of Geometric source preparation signal processing technique

Research paper thumbnail of Method and apparatus for image processing by generating probability distribution of images

Research paper thumbnail of Separation of Non-Stationary Natural Signals

CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of sourc... more CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of source separation use the assumption of statistical independence. To capture statistical independence higher order statistics are required. In this chapter we will demonstrate how higher ...

Research paper thumbnail of Convolutive Source Separation and Signal Modeling with ML

In independent component analysis (ICA) an instantaneous mix of sources can be recovered using ma... more In independent component analysis (ICA) an instantaneous mix of sources can be recovered using maximum Likelihood (ML). In Convolutive blind source separation (BSS) the mixture arises as a combination of differently convolved source signals due to time delays and a reverberating acoustic environment. Instead of modeling a particular time instant now a time window of the mixed signals has to be modeled. This allows to combine ICA with traditional ML signal modeling techniques. Here we use an auto-regressive (AR) model of the sources leading to a generalization of contextual ICA [18] to the convolutive case. This may improve source separation avoiding the typical whitening of the sources, and may allow us to incorporation simultaneous enhancing of the signal based on the AR models. 1 Introduction Independent component analysis (ICA) aims to find statistical independent signals in an instantaneous linear mix. This concept was first introduced and formalized by Comon [7]. In recent year...

Research paper thumbnail of On-line Blind Source Separation of Non-Stationary Signals

We have shown previously that non-stationarysignals recorded in a static multi-path environment c... more We have shown previously that non-stationarysignals recorded in a static multi-path environment can oftenbe recovered by simultaneously decorrelating varyingsecond order statistics. As typical sources are often moving,however, the multi-path channel is not static. We presenthere an on-line gradient algorithm with adaptive step sizein the frequency domain based on second derivatives, whichwe refer to as multiple adaptive decorrelation (MAD). Wecompared the

Research paper thumbnail of An Iterative Method for Vector Median Filtering

2007 IEEE International Conference on Image Processing, 2007

Research paper thumbnail of Apparatus for separating convolutively mixed signals derived from concurrent microphone recordings into plurality of sources using geometric information and adaptive beam forming

Research paper thumbnail of Separation of non-stationary natural signals

Principles and Practice, 2001

CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of sourc... more CiteSeerX - Document Details (Isaac Councill, Lee Giles): Most approaches to the problem of source separation use the assumption of statistical independence. To capture statistical independence higher order statistics are required. In this chapter we will demonstrate how higher ...

Research paper thumbnail of Convolutive blind separation of non-stationary sources

IEEE Transactions on Speech and Audio Processing, 2000

Log In