Robustness against external noise is an important requirement for automatic speech recognition (ASR) systems, when it comes to deploying them for practical applications. This thesis proposes and evaluates new feature-based approaches for improving the ASR noise robustness. These approaches are based on nonlinear transformations that, when applied to the spectrum or feature, aim to emphasize the...

In this paper, we introduce a new noise robust representation of speech signal obtained by locating points of potential importance in the spectrogram, and parameterizing the activity of time-frequency pattern around those points. These features are referred to as Spectro-Temporal Activity Pattern (STAP) features. The suitability of these features for noise robust speech recognition is examined ...

Multispectral remotely sensed images composed information over a large range of variation on frequencies (information) and these frequencies change over different regions (irregular or frequency variant behavior of the signal) which need to be estimated properly for an improved classification [1, 2, 3]. Multispectral remote sensing (RS) image data are basically complex in nature, which have bot...

hyperion hyperspectral data contains a very rich source of information from the earth surface that collects 242 narrow contiguous spectral bands. achieving this source of rich information is subject to the performance of suitable image processing methods on raw satellite data. satellite image processing methods can be classified into two categories of statistical-based and spectral-based. in th...

Numerical wave modeling in oceanic and coastal waters is usually based on a phase-averaged approach (spectral models), whereas close to shore, in the surf zone and in harbors, it is usually based on a phase-resolving approach (time domain models). Both approaches can be formulated in terms of the energy and phase spectrum of the waves. In the present project we are developing a model in which b...

This paper presents an investigation into the use of a 2-dimensional laser scanner (LiDAR) to obtain measurements of wave processes in the inner surf and swash zones of a microtidal beach (Rousty, Camargue, France). The bed is extracted at the wave-by-wave timescale using a variance threshold method on the time series. Individual wave properties were then retrieved from a local extrema analysis...

T his paper considers the imagery generated by multiple pendula swinging from a single beam. Although it is not generally detectable, the pendulum bobs lie on a sinusoid of contracting wavelength. Instead, an onlooker sees a sequence of diverse patterns produced by the bobs. Disparity between the undetected and perceived imagery is caused by insufficient signal sampling. The sequence of visual ...