Abstract: This paper describes a method for AWGN (Additive White Gaussian Noise) variance estimation in noisy stochastic signals, referred to as Multiplicative-Noising Variance Estimation (MNVE). The aim was to develop an estimation algorithm with minimal number of assumptions on the original signal structure. The provided MATLAB simulation and results analysis of the method applied on speech signals showed more accuracy than standardized AR (autoregressive) modeling noise estimation technique. In addition, great performance was observed on very low signal-to-noise ratios, which in general represents the worst case scenario for signal denoising methods. High execution time appears to be the only disadvantage of MNVE. After close examination of all the observed features of the proposed algorithm, it was concluded it is worth of exploring and that with some further adjustments and improvements can be enviably powerful.
Abstract: This paper aims at manipulating loop alignment in knitting a three-dimensional (3D) shape by its geometry. Two loop alignment methods are introduced to handle a surface with positive Gaussian curvature. As weft knitting is a two-dimensional (2D) knitting mechanism that the knitting cam carrying the feeders moves in two directions only, left and right, the knitted fabric generated grows in width and length but not in depth. Therefore, a 3D shape is required to be flattened to a 2D plane with surface area preserved for knitting. On this flattened plane, dimensional measurements are taken for loop alignment. The way these measurements being taken derived two different loop alignment methods. In this paper, only plain knitted structure was considered. Each knitted loop was taken as a basic unit for loop alignment in order to achieve the required geometric dimensions, without the inclusion of other stitches which give textural dimensions to the fabric. Two loop alignment methods were experimented and compared. Only one of these two can successfully preserve the dimensions of the shape.
Abstract: The present paper discusses the prediction of gas-liquid two-phase frictional pressure drop in a 2.12 mm horizontal circular minichannel using Artificial Neural Network (ANN). The experimental results are obtained with air as gas phase and water as liquid phase. The superficial gas velocity is kept in the range of 0.0236 m/s to 0.4722 m/s while the values of 0.0944 m/s, 0.1416 m/s and 0.1889 m/s are considered for superficial liquid velocity. The experimental results are predicted using different Artificial Neural Network (ANN) models. Networks used for prediction are radial basis, generalised regression, linear layer, cascade forward back propagation, feed forward back propagation, feed forward distributed time delay, layer recurrent, and Elman back propagation. Transfer functions used for networks are Linear (PURELIN), Logistic sigmoid (LOGSIG), tangent sigmoid (TANSIG) and Gaussian RBF. Combination of networks and transfer functions give different possible neural network models. These models are compared for Mean Absolute Relative Deviation (MARD) and Mean Relative Deviation (MRD) to identify the best predictive model of ANN.
Abstract: A physical model for guiding the wave in
photorefractive media is studied. Propagation of cos-Gaussian beam
as the special cases of sinusoidal-Gaussian beams in photorefractive
crystal is simulated numerically by the Crank-Nicolson method in
one dimension. Results show that the beam profile deforms as the
energy transfers from the center to the tails under propagation. This
simulation approach is of significant interest for application in optical
telecommunication. The results are presented graphically and
discussed.
Abstract: In this paper, approach to incoherent signal detection
in multi-element antenna array are researched and modeled. Two
types of useful signals with unknown wavefront were considered:
first one, deterministic (Barker code), and second one, random
(Gaussian distribution). The derivation of the sufficient statistics took
into account the linearity of the antenna array. The performance
characteristics and detecting curves are modeled and compared for
different useful signals parameters and for different number of
elements of the antenna array. Results of researches in case of some
additional conditions can be applied to a digital communications
systems.
Abstract: We present probabilistic multinomial Dirichlet
classification model for multidimensional data and Gaussian process
priors. Here, we have considered efficient computational method that
can be used to obtain the approximate posteriors for latent variables
and parameters needed to define the multiclass Gaussian process
classification model. We first investigated the process of inducing a
posterior distribution for various parameters and latent function by
using the variational Bayesian approximations and important sampling
method, and next we derived a predictive distribution of latent
function needed to classify new samples. The proposed model is
applied to classify the synthetic multivariate dataset in order to verify
the performance of our model. Experiment result shows that our model
is more accurate than the other approximation methods.
Abstract: In this paper, we propose the variational EM inference
algorithm for the multi-class Gaussian process classification model
that can be used in the field of human behavior recognition. This
algorithm can drive simultaneously both a posterior distribution of a
latent function and estimators of hyper-parameters in a Gaussian
process classification model with multiclass. Our algorithm is based
on the Laplace approximation (LA) technique and variational EM
framework. This is performed in two steps: called expectation and
maximization steps. First, in the expectation step, using the Bayesian
formula and LA technique, we derive approximately the posterior
distribution of the latent function indicating the possibility that each
observation belongs to a certain class in the Gaussian process
classification model. Second, in the maximization step, using a derived
posterior distribution of latent function, we compute the maximum
likelihood estimator for hyper-parameters of a covariance matrix
necessary to define prior distribution for latent function. These two
steps iteratively repeat until a convergence condition satisfies.
Moreover, we apply the proposed algorithm with human action
classification problem using a public database, namely, the KTH
human action data set. Experimental results reveal that the proposed
algorithm shows good performance on this data set.
Abstract: Magnetic Resonance Imaging (MRI) is one of the
most important medical imaging modality. Subjective assessment of
the image quality is regarded as the gold standard to evaluate MR
images. In this study, a database of 210 MR images which contains
ten reference images and 200 distorted images is presented. The
reference images were distorted with four types of distortions: Rician
Noise, Gaussian White Noise, Gaussian Blur and DCT compression.
The 210 images were assessed by ten subjects. The subjective scores
were presented in Difference Mean Opinion Score (DMOS). The
DMOS values were compared with four FR-IQA metrics. We have
used Pearson Linear Coefficient (PLCC) and Spearman Rank Order
Correlation Coefficient (SROCC) to validate the DMOS values. The
high correlation values of PLCC and SROCC shows that the DMOS
values are close to the objective FR-IQA metrics.
Abstract: The current-voltage (I-V) characteristics of Pd/n-GaN Schottky barrier were studied at temperatures over room temperature (300-470K). The values of ideality factor (n), zero-bias barrier height (φB0), flat barrier height (φBF) and series resistance (Rs) obtained from I-V-T measurements were found to be strongly temperature dependent while (φBo) increase, (n), (φBF) and (Rs) decrease with increasing temperature. The apparent Richardson constant was found to be 2.1x10-9 Acm-2K-2 and mean barrier height of 0.19 eV. After barrier height inhomogeneities correction, by assuming a Gaussian distribution (GD) of the barrier heights, the Richardson constant and the mean barrier height were obtained as 23 Acm-2K-2 and 1.78eV, respectively. The corrected Richardson constant was very closer to theoretical value of 26 Acm-2K-2.
Abstract: One of the most important challenging factors in
medical images is nominated as noise. Image denoising refers to the
improvement of a digital medical image that has been infected by
Additive White Gaussian Noise (AWGN). The digital medical image
or video can be affected by different types of noises. They are
impulse noise, Poisson noise and AWGN. Computed tomography
(CT) images are subjects to low quality due to the noise. Quality of
CT images is dependent on absorbed dose to patients directly in such
a way that increase in absorbed radiation, consequently absorbed
dose to patients (ADP), enhances the CT images quality. In this
manner, noise reduction techniques on purpose of images quality
enhancement exposing no excess radiation to patients is one the
challenging problems for CT images processing. In this work, noise
reduction in CT images was performed using two different
directional 2 dimensional (2D) transformations; i.e., Curvelet and
Contourlet and Discrete Wavelet Transform (DWT) thresholding
methods of BayesShrink and AdaptShrink, compared to each other
and we proposed a new threshold in wavelet domain for not only
noise reduction but also edge retaining, consequently the proposed
method retains the modified coefficients significantly that result good
visual quality. Data evaluations were accomplished by using two
criterions; namely, peak signal to noise ratio (PSNR) and Structure
similarity (Ssim).
Abstract: Speaker Identification (SI) is the task of establishing
identity of an individual based on his/her voice characteristics. The SI
task is typically achieved by two-stage signal processing: training and
testing. The training process calculates speaker specific feature
parameters from the speech and generates speaker models
accordingly. In the testing phase, speech samples from unknown
speakers are compared with the models and classified. Even though
performance of speaker identification systems has improved due to
recent advances in speech processing techniques, there is still need of
improvement. In this paper, a Closed-Set Tex-Independent Speaker
Identification System (CISI) based on a Multiple Classifier System
(MCS) is proposed, using Mel Frequency Cepstrum Coefficient
(MFCC) as feature extraction and suitable combination of vector
quantization (VQ) and Gaussian Mixture Model (GMM) together
with Expectation Maximization algorithm (EM) for speaker
modeling. The use of Voice Activity Detector (VAD) with a hybrid
approach based on Short Time Energy (STE) and Statistical
Modeling of Background Noise in the pre-processing step of the
feature extraction yields a better and more robust automatic speaker
identification system. Also investigation of Linde-Buzo-Gray (LBG)
clustering algorithm for initialization of GMM, for estimating the
underlying parameters, in the EM step improved the convergence rate
and systems performance. It also uses relative index as confidence
measures in case of contradiction in identification process by GMM
and VQ as well. Simulation results carried out on voxforge.org
speech database using MATLAB highlight the efficacy of the
proposed method compared to earlier work.
Abstract: In this paper, we propose moving object detection
method which is helpful for driver to safely take his/her car out of
parking lot. When moving objects such as motorbikes, pedestrians,
the other cars and some obstacles are detected at the rear-side of host
vehicle, the proposed algorithm can provide to driver warning. We
assume that the host vehicle is just before departure. Gaussian
Mixture Model (GMM) based background subtraction is basically
applied. Pre-processing such as smoothing and post-processing as
morphological filtering are added. We examine “which color space
has better performance for detection of moving objects?” Three color
spaces including RGB, YCbCr, and Y are applied and compared, in
terms of detection rate. Through simulation, we prove that RGB
space is more suitable for moving object detection based on
background subtraction.
Abstract: Any signal transmitted over a channel is corrupted by noise and interference. A host of channel coding techniques has been proposed to alleviate the effect of such noise and interference. Among these Turbo codes are recommended, because of increased capacity at higher transmission rates and superior performance over convolutional codes. The multimedia elements which are associated with ample amount of data are best protected by Turbo codes. Turbo decoder employs Maximum A-posteriori Probability (MAP) and Soft Output Viterbi Decoding (SOVA) algorithms. Conventional Turbo coded systems employ Equal Error Protection (EEP) in which the protection of all the data in an information message is uniform. Some applications involve Unequal Error Protection (UEP) in which the level of protection is higher for important information bits than that of other bits. In this work, enhancement to the traditional Log MAP decoding algorithm is being done by using optimized scaling factors for both the decoders. The error correcting performance in presence of UEP in Additive White Gaussian Noise channel (AWGN) and Rayleigh fading are analyzed for the transmission of image with Discrete Cosine Transform (DCT) as source coding technique. This paper compares the performance of log MAP, Modified log MAP (MlogMAP) and Enhanced log MAP (ElogMAP) algorithms used for image transmission. The MlogMAP algorithm is found to be best for lower Eb/N0 values but for higher Eb/N0 ElogMAP performs better with optimized scaling factors. The performance comparison of AWGN with fading channel indicates the robustness of the proposed algorithm. According to the performance of three different message classes, class3 would be more protected than other two classes. From the performance analysis, it is observed that ElogMAP algorithm with UEP is best for transmission of an image compared to Log MAP and MlogMAP decoding algorithms.
Abstract: Here, we study the characteristic feature of
conventional (ON-OFF keying) and soliton based transmission
system. We consider 20Gbps transmission system implemented with
Conventional Single Mode Fiber (C-SMF) to examine the role of
Gaussian pulse which is the characteristic of conventional
propagation and Hyperbolic-secant pulse which is the characteristic
of soliton propagation in it. We note the influence of these pulses
with respect to different dispersion lengths and soliton period in
conventional and soliton system respectively and evaluate the system
performance in terms of Quality factor. From the analysis, we could
prove that the soliton pulse has the consistent performance even for
long distance without dispersion compensation than the conventional
system as it is robust to dispersion. For the length of transmission of
200Km, soliton system yielded Q of 33.958 while the conventional
system totally exhausted with Q=0.
Abstract: A simple adaptive voice activity detector (VAD) is
implemented using Gabor and gammatone atomic decomposition of
speech for high Gaussian noise environments. Matching pursuit is
used for atomic decomposition, and is shown to achieve optimal
speech detection capability at high data compression rates for low
signal to noise ratios. The most active dictionary elements found by
matching pursuit are used for the signal reconstruction so that the
algorithm adapts to the individual speakers dominant time-frequency
characteristics. Speech has a high peak to average ratio enabling
matching pursuit greedy heuristic of highest inner products to isolate
high energy speech components in high noise environments. Gabor
and gammatone atoms are both investigated with identical
logarithmically spaced center frequencies, and similar bandwidths.
The algorithm performs equally well for both Gabor and gammatone
atoms with no significant statistical differences. The algorithm
achieves 70% accuracy at a 0 dB SNR, 90% accuracy at a 5 dB SNR
and 98% accuracy at a 20dB SNR using 30d B SNR as a reference
for voice activity.
Abstract: Image enhancement is a challenging issue in many applications. In the last two decades, there are various filters developed. This paper proposes a novel method which removes Gaussian noise from the gray scale images. The proposed technique is compared with Enhanced Fuzzy Peer Group Filter (EFPGF) for various noise levels. Experimental results proved that the proposed filter achieves better Peak-Signal-to-Noise-Ratio PSNR than the existing techniques. The proposed technique achieves 1.736dB gain in PSNR than the EFPGF technique.
Abstract: A cyclostationary Gaussian linearization method is
formulated for investigating the time average response of nonlinear
system under sinusoidal signal and white noise excitation. The
quantitative measure of cyclostationary mean, variance, spectrum of
mean amplitude, and mean power spectral density of noise are
analyzed. The qualitative response behavior of stochastic jump and
bifurcation are investigated. The validity of the present approach in
predicting the quantitative and qualitative statistical responses is
supported by utilizing Monte Carlo simulations. The present analysis
without imposing restrictive analytical conditions can be directly
derived by solving non-linear algebraic equations. The analytical
solution gives reliable quantitative and qualitative prediction of mean
and noise response for the Duffing system subjected to both sinusoidal
signal and white noise excitation.
Abstract: The detection of moving objects from a video image
sequences is very important for object tracking, activity recognition,
and behavior understanding in video surveillance.
The most used approach for moving objects detection / tracking is
background subtraction algorithms. Many approaches have been
suggested for background subtraction. But, these are illumination
change sensitive and the solutions proposed to bypass this problem
are time consuming.
In this paper, we propose a robust yet computationally efficient
background subtraction approach and, mainly, focus on the ability to
detect moving objects on dynamic scenes, for possible applications in
complex and restricted access areas monitoring, where moving and
motionless persons must be reliably detected. It consists of three
main phases, establishing illumination changes invariance,
background/foreground modeling and morphological analysis for
noise removing.
We handle illumination changes using Contrast Limited Histogram
Equalization (CLAHE), which limits the intensity of each pixel to
user determined maximum. Thus, it mitigates the degradation due to
scene illumination changes and improves the visibility of the video
signal. Initially, the background and foreground images are extracted
from the video sequence. Then, the background and foreground
images are separately enhanced by applying CLAHE.
In order to form multi-modal backgrounds we model each channel
of a pixel as a mixture of K Gaussians (K=5) using Gaussian Mixture
Model (GMM). Finally, we post process the resulting binary
foreground mask using morphological erosion and dilation
transformations to remove possible noise.
For experimental test, we used a standard dataset to challenge the
efficiency and accuracy of the proposed method on a diverse set of
dynamic scenes.
Abstract: This paper presents two techniques, local feature
extraction using image spectrum and low frequency spectrum
modelling using GMM to capture the underlying statistical
information to improve the performance of face recognition
system. Local spectrum features are extracted using overlap sub
block window that are mapped on the face image. For each of this
block, spatial domain is transformed to frequency domain using
DFT. A low frequency coefficient is preserved by discarding high
frequency coefficients by applying rectangular mask on the
spectrum of the facial image. Low frequency information is non-
Gaussian in the feature space and by using combination of several
Gaussian functions that has different statistical properties, the best
feature representation can be modelled using probability density
function. The recognition process is performed using maximum
likelihood value computed using pre-calculated GMM components.
The method is tested using FERET datasets and is able to achieved
92% recognition rates.
Abstract: Over the past few years, the online multimedia
collection has grown at a fast pace. Several companies showed
interest to study the different ways to organise the amount of audio
information without the need of human intervention to generate
metadata. In the past few years, many applications have emerged on
the market which are capable of identifying a piece of music in a
short time. Different audio effects and degradation make it much
harder to identify the unknown piece. In this paper, an audio
fingerprinting system which makes use of a non-parametric based
algorithm is presented. Parametric analysis is also performed using
Gaussian Mixture Models (GMMs). The feature extraction methods
employed are the Mel Spectrum Coefficients and the MPEG-7 basic
descriptors. Bin numbers replaced the extracted feature coefficients
during the non-parametric modelling. The results show that nonparametric
analysis offer potential results as the ones mentioned in
the literature.