We propose a simultaneous spatio-temporal filter optimization algorithm for the single trial ElectroEncephaloGraphy (EEG) classification problem. The algorithm is a generalization of the Common Spatial Pattern (CSP) algorithm, which incorporates non-homogeneous weighting of the cross-spectrum matrices. We alternately update the spectral weighting coefficients and the spatial projection. The cro...

A filter structure is proposed which has a significantly lower signal delay than common frequency-domain filtering by means of an analysis-synthesis filter-bank (AS FB). The recently published filter-bank equalizer (FBE) is a filter(-bank) with a lower signal delay compared to the corresponding AS FB. In this contribution, the signal delay of the FBE is further reduced by approximating the time...

Robustness for BCI is usually obtained at the classifier level Non-stationarity is inherent to brain signals We propose to leverage this at the feature extraction level

This paper proposes a new fusion method for multiespectral (MULTI) and panchromatic (PAN) images that uses a highly anisotropic and redundant representation of images. This methodology join the simplicity of the Wavelet transform, calculated using the à trous algorithm, with the benefits of multidirectional transforms like Contourlet Transform. That has permitted an adequate extraction of infor...

Major depressive disorder (MDD) has become a leading contributor to the global burden of disease; however, there are currently no reliable biological markers or physiological measurements for efficiently and effectively dissecting the heterogeneity of MDD. Here we propose a novel method based on scalp electroencephalography (EEG) signals and a robust spectral-spatial EEG feature extractor calle...

We introduce a novel, adaptive image representation using spatially varying wavelet packets (WPs), Our adaptive representation uses the fast double-tree algorithm introduced previously (Herley et al., 1993) to optimize an operational rate-distortion (R-D) cost function, as is appropriate for the lossy image compression framework. This involves jointly determining which filter bank tree (WP freq...

In smart antenna system (SAS), the most compelling research work is the space-division multiple-access (SDMA) which includes uplink source separation and downlink selective transmission. In this paper, we propose a subspace-based spatial signature estimation algorithm of cochannel multiuser signals by jointly exploiting cumulant and cyclostationarity, and apply it to multiuser signal copy in up...

This paper presents a method for classification of Mammographic Microcalcifications using Quadmture Mirror Filter (QMF) bank subband decomposition as a tool for extracting textuml features. The QMF bank splits the input image into four subbands: low-low band, low-high band, high-low band and high-high band, and textural features baaed on eo-occurrence matrices are computed from the subsampled s...

The restoration of an entire epicortical ECoG-potential pattern from spatially sampled EEG-potentials can be performed according to the general principle of inverse spatial filtering: if a forward spatial filter is an averaging one, then a mutually inverse spatial filter can form a sharpened pattern from a slurred sampled pattern thus restoring more or less a spatial contrast of a primary sourc...