Abstract: Matching algorithms have significant importance in
speaker recognition. Feature vectors of the unknown utterance are
compared to feature vectors of the modeled speakers as a last step in
speaker recognition. A similarity score is found for every model in
the speaker database. Depending on the type of speaker recognition,
these scores are used to determine the author of unknown speech
samples. For speaker verification, similarity score is tested against a
predefined threshold and either acceptance or rejection result is
obtained. In the case of speaker identification, the result depends on
whether the identification is open set or closed set. In closed set
identification, the model that yields the best similarity score is
accepted. In open set identification, the best score is tested against a
threshold, so there is one more possible output satisfying the
condition that the speaker is not one of the registered speakers in
existing database. This paper focuses on closed set speaker
identification using a modified version of a well known matching
algorithm. The results of new matching algorithm indicated better
performance on YOHO international speaker recognition database.
Abstract: The wavelet transform is one of the most important
method used in signal processing. In this study, we have introduced
frequency-energy characteristics of local earthquakes using discrete
wavelet transform. Frequency-energy characteristic was analyzed
depend on difference between P and S wave arrival time and noise
within records. We have found that local earthquakes have similar
characteristics. If frequency-energy characteristics can be found
accurately, this gives us a hint to calculate P and S wave arrival time.
It can be seen that wavelet transform provides successful
approximation for this. In this study, 100 earthquakes with 500
records were analyzed approximately.
Abstract: This article is devoted to the numerical solution of
large-scale quadratic eigenvalue problems. Such problems arise in
a wide variety of applications, such as the dynamic analysis of
structural mechanical systems, acoustic systems, fluid mechanics,
and signal processing. We first introduce a generalized second-order
Krylov subspace based on a pair of square matrices and two initial
vectors and present a generalized second-order Arnoldi process for
constructing an orthonormal basis of the generalized second-order
Krylov subspace. Then, by using the projection technique and the
refined projection technique, we propose a restarted generalized
second-order Arnoldi method and a restarted refined generalized
second-order Arnoldi method for computing some eigenpairs of largescale
quadratic eigenvalue problems. Some theoretical results are also
presented. Some numerical examples are presented to illustrate the
effectiveness of the proposed methods.
Abstract: This paper proposes an architecture of dynamically
reconfigurable arithmetic circuit. Dynamic reconfiguration is a
technique to realize required functions by changing hardware
construction during operations. The proposed circuit is based on a
complex number multiply-accumulation circuit which is used
frequently in the field of digital signal processing. In addition, the
proposed circuit performs real number double precision arithmetic
operations. The data formats are single and double precision floating
point number based on IEEE754. The proposed circuit is designed
using VHDL, and verified the correct operation by simulations and
experiments.
Abstract: Motor imagery classification provides an important basis for designing Brain Machine Interfaces [BMI]. A BMI captures and decodes brain EEG signals and transforms human thought into actions. The ability of an individual to control his EEG through imaginary mental tasks enables him to control devices through the BMI. This paper presents a method to design a four state BMI using EEG signals recorded from the C3 and C4 locations. Principle features extracted through principle component analysis of the segmented EEG are analyzed using two novel classification algorithms using Elman recurrent neural network and functional link neural network. Performance of both classifiers is evaluated using a particle swarm optimization training algorithm; results are also compared with the conventional back propagation training algorithm. EEG motor imagery recorded from two subjects is used in the offline analysis. From overall classification performance it is observed that the BP algorithm has higher average classification of 93.5%, while the PSO algorithm has better training time and maximum classification. The proposed methods promises to provide a useful alternative general procedure for motor imagery classification
Abstract: In this paper, we propose a fully-utilized, block-based 2D DWT (discrete wavelet transform) architecture, which consists of four 1D DWT filters with two-channel QMF lattice structure. The proposed architecture requires about 2MN-3N registers to save the intermediate results for higher level decomposition, where M and N stand for the filter length and the row width of the image respectively. Furthermore, the proposed 2D DWT processes in horizontal and vertical directions simultaneously without an idle period, so that it computes the DWT for an N×N image in a period of N2(1-2-2J)/3. Compared to the existing approaches, the proposed architecture shows 100% of hardware utilization and high throughput rates. To mitigate the long critical path delay due to the cascaded lattices, we can apply the pipeline technique with four stages, while retaining 100% of hardware utilization. The proposed architecture can be applied in real-time video signal processing.
Abstract: The whole work is based on possibility to use Lego Mindstorms robotics systems to reduce costs. Lego Mindstorms consists of a wide variety of hardware components necessary to simulate, programme and test of robotics systems in practice. To programme algorithm, which simulates space using the ultrasonic sensor, was used development environment supplied with kit. Software Matlab was used to render values afterwards they were measured by ultrasonic sensor. The algorithm created for this paper uses theoretical knowledge from area of signal processing. Data being processed by algorithm are collected by ultrasonic sensor that scans 2D space in front of it. Ultrasonic sensor is placed on moving arm of robot which provides horizontal moving of sensor. Vertical movement of sensor is provided by wheel drive. The robot follows map in order to get correct positioning of measured data. Based on discovered facts it is possible to consider Lego Mindstorm for low-cost and capable kit for real-time modelling.
Abstract: In this paper, acoustic techniques are used to detect hidden insect infestations of date palm tress (Phoenix dactylifera L.). In particular, we use an acoustic instrument for early discovery of the presence of a destructive insect pest commonly known as the Red Date Palm Weevil (RDPW) and scientifically as Rhynchophorus ferrugineus (Olivier). This type of insect attacks date palm tress and causes irreversible damages at late stages. As a result, the infected trees must be destroyed. Therefore, early presence detection is a major part in controlling the spread and economic damage caused by this type of infestation. Furthermore monitoring and early detection of the disease can asses in taking appropriate measures such as isolating or treating the infected trees. The acoustic system is evaluated in terms of its ability for early discovery of hidden bests inside the tested tree. When signal acquisitions is completed for a number of date palms, a signal processing technique known as time-frequency analysis is evaluated in terms of providing an estimate that can be visually used to recognize the acoustic signature of the RDPW. The testing instrument was tested in the laboratory first then; it was used on suspected or infested tress in the field. The final results indicate that the acoustic monitoring approach along with signal processing techniques are very promising for the early detection of presence of the larva as well as the adult pest in the date palms.
Abstract: Acoustic Imaging based sound localization using microphone
array is a challenging task in digital-signal processing.
Discrete Fourier transform (DFT) based near-field acoustical holography
(NAH) is an important acoustical technique for sound source
localization and provide an efficient solution to the ill-posed problem.
However, in practice, due to the usage of small curtailed aperture
and its consequence of significant spectral leakage, the DFT could
not reconstruct the active-region-of-sound (AROS) effectively, especially
near the edges of aperture. In this paper, we emphasize the
fundamental problems of DFT-based NAH, provide a solution to
spectral leakage effect by the extrapolation based on linear predictive
coding and 2D Tukey windowing. This approach has been tested to
localize the single and multi-point sound sources. We observe that
incorporating extrapolation technique increases the spatial resolution,
localization accuracy and reduces spectral leakage when small curtail
aperture with a lower number of sensors accounts.
Abstract: The goal of this work is to improve the efficiency and the reliability of the automatic artifact rejection, in particular from the Electroencephalographic (EEG) recordings. Artifact rejection is a key topic in signal processing. The artifacts are unwelcome signals that may occur during the signal acquisition and that may alter the analysis of the signals themselves. A technique for the automatic artifact rejection, based on the Independent Component Analysis (ICA) for the artifact extraction and on some high order statistics such as kurtosis and Shannon-s entropy, was proposed some years ago in literature. In this paper we enhance this technique introducing the Renyi-s entropy. The performance of our method was tested exploiting the Independent Component scalp maps and it was compared to the performance of the method in literature and it showed to outperform it.
Abstract: Stochastic resonance (SR) is a phenomenon whereby
the signal transmission or signal processing through certain nonlinear
systems can be improved by adding noise. This paper discusses SR in
nonlinear signal detection by a simple test statistic, which can be
computed from multiple noisy data in a binary decision problem based
on a maximum a posteriori probability criterion. The performance of
detection is assessed by the probability of detection error Per . When
the input signal is subthreshold signal, we establish that benefit from
noise can be gained for different noises and confirm further that the
subthreshold SR exists in nonlinear signal detection. The efficacy of
SR is significantly improved and the minimum of Per can
dramatically approach to zero as the sample number increases. These
results show the robustness of SR in signal detection and extend the
applicability of SR in signal processing.
Abstract: A complex valued neural network is a neural network
which consists of complex valued input and/or weights and/or thresholds
and/or activation functions. Complex-valued neural networks
have been widening the scope of applications not only in electronics
and informatics, but also in social systems. One of the most important
applications of the complex valued neural network is in signal
processing. In Neural networks, generalized mean neuron model
(GMN) is often discussed and studied. The GMN includes a new
aggregation function based on the concept of generalized mean of all
the inputs to the neuron. This paper aims to present exhaustive results
of using Generalized Mean Neuron model in a complex-valued neural
network model that uses the back-propagation algorithm (called
-Complex-BP-) for learning. Our experiments results demonstrate the
effectiveness of a Generalized Mean Neuron Model in a complex
plane for signal processing over a real valued neural network. We
have studied and stated various observations like effect of learning
rates, ranges of the initial weights randomly selected, error functions
used and number of iterations for the convergence of error required on
a Generalized Mean neural network model. Some inherent properties
of this complex back propagation algorithm are also studied and
discussed.
Abstract: The paper presents a novel idea to control computer
mouse cursor movement with human eyes. In this paper, a working
of the product has been described as to how it helps the special
people share their knowledge with the world. Number of traditional
techniques such as Head and Eye Movement Tracking Systems etc.
exist for cursor control by making use of image processing in which
light is the primary source. Electro-oculography (EOG) is a new
technology to sense eye signals with which the mouse cursor can be
controlled. The signals captured using sensors, are first amplified,
then noise is removed and then digitized, before being transferred to
PC for software interfacing.
Abstract: This paper describes how the correct endian mode of
the TMS320C6713 DSK board can be identified. It also explains how
the TMS320C6713 DSK board can be used in the little endian and in
the big endian modes for assembly language programming in
particular and for signal processing in general. Similarly, it discusses
how crucially important it is for a user of the TMS320C6713 DSK
board to identify the mode of operation and then use it correctly
during the development stages of the assembly language
programming; otherwise, it will cause unnecessary confusion and
erroneous results as far as storing data into the memory and loading
data from the memory is concerned. Furthermore, it highlights and
strongly recommends to the users of the TMS320C6713 DSK board
to be aware of the availability and importance of various display
options in the Code Composer Studio (CCS) for correctly
interpreting and displaying the desired data in the memory. The
information presented in this paper will be of great importance and
interest to those practitioners and developers who wants to use the
TMS320C6713 DSK board for assembly language programming as
well as input-output signal processing manipulations. Finally,
examples that clearly illustrate the concept are presented.
Abstract: Heart failure is the most common reason of death
nowadays, but if the medical help is given directly, the patient-s life
may be saved in many cases. Numerous heart diseases can be
detected by means of analyzing electrocardiograms (ECG). Artificial
Neural Networks (ANN) are computer-based expert systems that
have proved to be useful in pattern recognition tasks. ANN can be
used in different phases of the decision-making process, from
classification to diagnostic procedures. This work concentrates on a
review followed by a novel method.
The purpose of the review is to assess the evidence of healthcare
benefits involving the application of artificial neural networks to the
clinical functions of diagnosis, prognosis and survival analysis, in
ECG signals. The developed method is based on a compound neural
network (CNN), to classify ECGs as normal or carrying an
AtrioVentricular heart Block (AVB). This method uses three
different feed forward multilayer neural networks. A single output
unit encodes the probability of AVB occurrences. A value between 0
and 0.1 is the desired output for a normal ECG; a value between 0.1
and 1 would infer an occurrence of an AVB. The results show that
this compound network has a good performance in detecting AVBs,
with a sensitivity of 90.7% and a specificity of 86.05%. The accuracy
value is 87.9%.