Abstract: In this work, we present for the first time in our
perception an efficient digital watermarking scheme for mpeg audio
layer 3 files that operates directly in the compressed data domain,
while manipulating the time and subband/channel domain. In
addition, it does not need the original signal to detect the watermark.
Our scheme was implemented taking special care for the efficient
usage of the two limited resources of computer systems: time and
space. It offers to the industrial user the capability of watermark
embedding and detection in time immediately comparable to the real
music time of the original audio file that depends on the mpeg
compression, while the end user/audience does not face any artifacts
or delays hearing the watermarked audio file. Furthermore, it
overcomes the disadvantage of algorithms operating in the PCMData
domain to be vulnerable to compression/recompression attacks,
as it places the watermark in the scale factors domain and not in the
digitized sound audio data. The strength of our scheme, that allows it
to be used with success in both authentication and copyright
protection, relies on the fact that it gives to the users the enhanced
capability their ownership of the audio file not to be accomplished
simply by detecting the bit pattern that comprises the watermark
itself, but by showing that the legal owner knows a hard to compute
property of the watermark.
Abstract: Using Dynamic Bayesian Networks (DBN) to model genetic regulatory networks from gene expression data is one of the major paradigms for inferring the interactions among genes. Averaging a collection of models for predicting network is desired, rather than relying on a single high scoring model. In this paper, two kinds of model searching approaches are compared, which are Greedy hill-climbing Search with Restarts (GSR) and Markov Chain Monte Carlo (MCMC) methods. The GSR is preferred in many papers, but there is no such comparison study about which one is better for DBN models. Different types of experiments have been carried out to try to give a benchmark test to these approaches. Our experimental results demonstrated that on average the MCMC methods outperform the GSR in accuracy of predicted network, and having the comparable performance in time efficiency. By proposing the different variations of MCMC and employing simulated annealing strategy, the MCMC methods become more efficient and stable. Apart from comparisons between these approaches, another objective of this study is to investigate the feasibility of using DBN modeling approaches for inferring gene networks from few snapshots of high dimensional gene profiles. Through synthetic data experiments as well as systematic data experiments, the experimental results revealed how the performances of these approaches can be influenced as the target gene network varies in the network size, data size, as well as system complexity.
Abstract: In this paper we address the issue of classifying the fluorescent intensity of a sample in Indirect Immuno-Fluorescence (IIF). Since IIF is a subjective, semi-quantitative test in its very nature, we discuss a strategy to reliably label the image data set by using the diagnoses performed by different physicians. Then, we discuss image pre-processing, feature extraction and selection. Finally, we propose two ANN-based classifiers that can separate intrinsically dubious samples and whose error tolerance can be flexibly set. Measured performance shows error rates less than 1%, which candidates the method to be used in daily medical practice either to perform pre-selection of cases to be examined, or to act as a second reader.
Abstract: In developing a text-to-speech system, it is well
known that the accuracy of information extracted from a text is
crucial to produce high quality synthesized speech. In this paper, a
new scheme for converting text into its equivalent phonetic spelling
is introduced and developed. This method is applicable to many
applications in text to speech converting systems and has many
advantages over other methods. The proposed method can also
complement the other methods with a purpose of improving their
performance. The proposed method is a probabilistic model and is
based on Smooth Ergodic Hidden Markov Model. This model can be
considered as an extension to HMM. The proposed method is applied
to Persian language and its accuracy in converting text to speech
phonetics is evaluated using simulations.
Abstract: Use of the Internet and the World-Wide-Web
(WWW) has become widespread in recent years and mobile agent
technology has proliferated at an equally rapid rate. In this scenario
load balancing becomes important for P2P systems. Beside P2P
systems can be highly heterogeneous, i.e., they may consists of peers
that range from old desktops to powerful servers connected to
internet through high-bandwidth lines. There are various loads
balancing policies came into picture. Primitive one is Message
Passing Interface (MPI). Its wide availability and portability make it
an attractive choice; however the communication requirements are
sometimes inefficient when implementing the primitives provided by
MPI. In this scenario we use the concept of mobile agent because
Mobile agent (MA) based approach have the merits of high
flexibility, efficiency, low network traffic, less communication
latency as well as highly asynchronous. In this study we present
decentralized load balancing scheme using mobile agent technology
in which when a node is overloaded, task migrates to less utilized
nodes so as to share the workload. However, the decision of which
nodes receive migrating task is made in real-time by defining certain
load balancing policies. These policies are executed on PMADE (A
Platform for Mobile Agent Distribution and Execution) in
decentralized manner using JuxtaNet and various load balancing
metrics are discussed.
Abstract: In this paper we investigated a number of the Internet
congestion control algorithms that has been developed in the last few
years. It was obviously found that many of these algorithms were
designed to deal with the Internet traffic merely as a train of
consequent packets. Other few algorithms were specifically tailored
to handle the Internet congestion caused by running media traffic that
represents audiovisual content. This later set of algorithms is
considered to be aware of the nature of this media content. In this
context we briefly explained a number of congestion control
algorithms and hence categorized them into the two following
categories: i) Media congestion control algorithms. ii) Common
congestion control algorithms. We hereby recommend the usage of
the media congestion control algorithms for the reason of being
media content-aware rather than the other common type of
algorithms that blindly manipulates such traffic. We showed that the
spread of such media content-aware algorithms over Internet will
lead to better congestion control status in the coming years. This is
due to the observed emergence of the era of digital convergence
where the media traffic type will form the majority of the Internet
traffic.
Abstract: The permanent magnet synchronous motor (PMSM) is
very useful in many applications. Vector control of PMSM is popular
kind of its control. In this paper, at first an optimal vector control for
PMSM is designed and then results are compared with conventional
vector control. Then, it is assumed that the measurements are noisy
and linear quadratic Gaussian (LQG) methodology is used to filter
the noises. The results of noisy optimal vector control and filtered
optimal vector control are compared to each other. Nonlinearity of
PMSM and existence of inverter in its control circuit caused that the
system is nonlinear and time-variant. With deriving average model,
the system is changed to nonlinear time-invariant and then the
nonlinear system is converted to linear system by linearization of
model around average values. This model is used to optimize vector
control then two optimal vector controls are compared to each other.
Simulation results show that the performance and robustness to noise
of the control system has been highly improved.
Abstract: Identity verification of authentic persons by their multiview faces is a real valued problem in machine vision. Multiview faces are having difficulties due to non-linear representation in the feature space. This paper illustrates the usability of the generalization of LDA in the form of canonical covariate for face recognition to multiview faces. In the proposed work, the Gabor filter bank is used to extract facial features that characterized by spatial frequency, spatial locality and orientation. Gabor face representation captures substantial amount of variations of the face instances that often occurs due to illumination, pose and facial expression changes. Convolution of Gabor filter bank to face images of rotated profile views produce Gabor faces with high dimensional features vectors. Canonical covariate is then used to Gabor faces to reduce the high dimensional feature spaces into low dimensional subspaces. Finally, support vector machines are trained with canonical sub-spaces that contain reduced set of features and perform recognition task. The proposed system is evaluated with UMIST face database. The experiment results demonstrate the efficiency and robustness of the proposed system with high recognition rates.
Abstract: One of the popular methods for recognition of facial
expressions such as happiness, sadness and surprise is based on
deformation of facial features. Motion vectors which show these
deformations can be specified by the optical flow. In this method, for
detecting emotions, the resulted set of motion vectors are compared
with standard deformation template that caused by facial expressions.
In this paper, a new method is introduced to compute the quantity of
likeness in order to make decision based on the importance of
obtained vectors from an optical flow approach. For finding the
vectors, one of the efficient optical flow method developed by
Gautama and VanHulle[17] is used. The suggested method has been
examined over Cohn-Kanade AU-Coded Facial Expression Database,
one of the most comprehensive collections of test images available.
The experimental results show that our method could correctly
recognize the facial expressions in 94% of case studies. The results
also show that only a few number of image frames (three frames) are
sufficient to detect facial expressions with rate of success of about
83.3%. This is a significant improvement over the available methods.
Abstract: Segmentation is an important step in medical image
analysis and classification for radiological evaluation or computer
aided diagnosis. This paper presents the problem of inaccurate lung
segmentation as observed in algorithms presented by researchers
working in the area of medical image analysis. The different lung
segmentation techniques have been tested using the dataset of 19
patients consisting of a total of 917 images. We obtained datasets of
11 patients from Ackron University, USA and of 8 patients from
AGA Khan Medical University, Pakistan. After testing the algorithms
against datasets, the deficiencies of each algorithm have been
highlighted.
Abstract: Several works regarding facial recognition have dealt with methods which identify isolated characteristics of the face or with templates which encompass several regions of it. In this paper a new technique which approaches the problem holistically dispensing with the need to identify geometrical characteristics or regions of the face is introduced. The characterization of a face is achieved by randomly sampling selected attributes of the pixels of its image. From this information we construct a set of data, which correspond to the values of low frequencies, gradient, entropy and another several characteristics of pixel of the image. Generating a set of “p" variables. The multivariate data set with different polynomials minimizing the data fitness error in the minimax sense (L∞ - Norm) is approximated. With the use of a Genetic Algorithm (GA) it is able to circumvent the problem of dimensionality inherent to higher degree polynomial approximations. The GA yields the degree and values of a set of coefficients of the polynomials approximating of the image of a face. By finding a family of characteristic polynomials from several variables (pixel characteristics) for each face (say Fi ) in the data base through a resampling process the system in use, is trained. A face (say F ) is recognized by finding its characteristic polynomials and using an AdaBoost Classifier from F -s polynomials to each of the Fi -s polynomials. The winner is the polynomial family closer to F -s corresponding to target face in data base.
Abstract: Data stream analysis is the process of computing
various summaries and derived values from large amounts of data
which are continuously generated at a rapid rate. The nature of a
stream does not allow a revisit on each data element. Furthermore,
data processing must be fast to produce timely analysis results. These
requirements impose constraints on the design of the algorithms to
balance correctness against timely responses. Several techniques
have been proposed over the past few years to address these
challenges. These techniques can be categorized as either dataoriented
or task-oriented. The data-oriented approach analyzes a
subset of data or a smaller transformed representation, whereas taskoriented
scheme solves the problem directly via approximation
techniques. We propose a hybrid approach to tackle the data stream
analysis problem. The data stream has been both statistically
transformed to a smaller size and computationally approximated its
characteristics. We adopt a Monte Carlo method in the approximation
step. The data reduction has been performed horizontally and
vertically through our EMR sampling method. The proposed method
is analyzed by a series of experiments. We apply our algorithm on
clustering and classification tasks to evaluate the utility of our
approach.
Abstract: In this paper we present a new method for coin
identification. The proposed method adopts a hybrid scheme using
Eigenvalues of covariance matrix, Circular Hough Transform (CHT)
and Bresenham-s circle algorithm. The statistical and geometrical
properties of the small and large Eigenvalues of the covariance
matrix of a set of edge pixels over a connected region of support are
explored for the purpose of circular object detection. Sparse matrix
technique is used to perform CHT. Since sparse matrices squeeze
zero elements and contain only a small number of non-zero elements,
they provide an advantage of matrix storage space and computational
time. Neighborhood suppression scheme is used to find the valid
Hough peaks. The accurate position of the circumference pixels is
identified using Raster scan algorithm which uses geometrical
symmetry property. After finding circular objects, the proposed
method uses the texture on the surface of the coins called texton,
which are unique properties of coins, refers to the fundamental micro
structure in generic natural images. This method has been tested on
several real world images including coin and non-coin images. The
performance is also evaluated based on the noise withstanding
capability.
Abstract: This paper proposes view-point insensitive human
pose recognition system using neural network. Recognition system
consists of silhouette image capturing module, data driven database,
and neural network. The advantages of our system are first, it is
possible to capture multiple view-point silhouette images of 3D human
model automatically. This automatic capture module is helpful to
reduce time consuming task of database construction. Second, we
develop huge feature database to offer view-point insensitivity at pose
recognition. Third, we use neural network to recognize human pose
from multiple-view because every pose from each model have similar
feature patterns, even though each model has different appearance and
view-point. To construct database, we need to create 3D human model
using 3D manipulate tools. Contour shape is used to convert silhouette
image to feature vector of 12 degree. This extraction task is processed
semi-automatically, which benefits in that capturing images and
converting to silhouette images from the real capturing environment is
needless. We demonstrate the effectiveness of our approach with
experiments on virtual environment.
Abstract: Finger spelling is an art of communicating by signs
made with fingers, and has been introduced into sign language to serve
as a bridge between the sign language and the verbal language.
Previous approaches to finger spelling recognition are classified into
two categories: glove-based and vision-based approaches. The
glove-based approach is simpler and more accurate recognizing work
of hand posture than vision-based, yet the interfaces require the user to
wear a cumbersome and carry a load of cables that connected the
device to a computer. In contrast, the vision-based approaches provide
an attractive alternative to the cumbersome interface, and promise
more natural and unobtrusive human-computer interaction. The
vision-based approaches generally consist of two steps: hand
extraction and recognition, and two steps are processed independently.
This paper proposes real-time vision-based Korean finger spelling
recognition system by integrating hand extraction into recognition.
First, we tentatively detect a hand region using CAMShift algorithm.
Then fill factor and aspect ratio estimated by width and height
estimated by CAMShift are used to choose candidate from database,
which can reduce the number of matching in recognition step. To
recognize the finger spelling, we use DTW(dynamic time warping)
based on modified chain codes, to be robust to scale and orientation
variations. In this procedure, since accurate hand regions, without
holes and noises, should be extracted to improve the precision, we use
graph cuts algorithm that globally minimize the energy function
elegantly expressed by Markov random fields (MRFs). In the
experiments, the computational times are less than 130ms, and the
times are not related to the number of templates of finger spellings in
database, as candidate templates are selected in extraction step.
Abstract: In this paper, we propose an improved 3D star skeleton
technique, which is a suitable skeletonization for human posture representation
and reflects the 3D information of human posture.
Moreover, the proposed technique is simple and then can be performed
in real-time. The existing skeleton construction techniques, such as
distance transformation, Voronoi diagram, and thinning, focus on the
precision of skeleton information. Therefore, those techniques are not
applicable to real-time posture recognition since they are computationally
expensive and highly susceptible to noise of boundary. Although
a 2D star skeleton was proposed to complement these problems,
it also has some limitations to describe the 3D information of the
posture. To represent human posture effectively, the constructed skeleton
should consider the 3D information of posture. The proposed 3D
star skeleton contains 3D data of human, and focuses on human action
and posture recognition. Our 3D star skeleton uses the 8 projection
maps which have 2D silhouette information and depth data of human
surface. And the extremal points can be extracted as the features of 3D
star skeleton, without searching whole boundary of object. Therefore,
on execution time, our 3D star skeleton is faster than the “greedy" 3D
star skeleton using the whole boundary points on the surface. Moreover,
our method can offer more accurate skeleton of posture than the
existing star skeleton since the 3D data for the object is concerned.
Additionally, we make a codebook, a collection of representative 3D
star skeletons about 7 postures, to recognize what posture of constructed
skeleton is.
Abstract: Multiple sequence alignment is a fundamental part in
many bioinformatics applications such as phylogenetic analysis.
Many alignment methods have been proposed. Each method gives a
different result for the same data set, and consequently generates a
different phylogenetic tree. Hence, the chosen alignment method
affects the resulting tree. However in the literature, there is no
evaluation of multiple alignment methods based on the comparison of
their phylogenetic trees. This work evaluates the following eight
aligners: ClustalX, T-Coffee, SAGA, MUSCLE, MAFFT, DIALIGN,
ProbCons and Align-m, based on their phylogenetic trees (test trees)
produced on a given data set. The Neighbor-Joining method is used
to estimate trees. Three criteria, namely, the dNNI, the dRF and the
Id_Tree are established to test the ability of different alignment
methods to produce closer test tree compared to the reference one
(true tree). Results show that the method which produces the most
accurate alignment gives the nearest test tree to the reference tree.
MUSCLE outperforms all aligners with respect to the three criteria
and for all datasets, performing particularly better when sequence
identities are within 10-20%. It is followed by T-Coffee at lower
sequence identity (30%), trees scores of all methods
become similar.
Abstract: For the communication between human and computer
in an interactive computing environment, the gesture recognition is
studied vigorously. Therefore, a lot of studies have proposed efficient
methods about the recognition algorithm using 2D camera captured
images. However, there is a limitation to these methods, such as the
extracted features cannot fully represent the object in real world.
Although many studies used 3D features instead of 2D features for
more accurate gesture recognition, the problem, such as the processing
time to generate 3D objects, is still unsolved in related researches.
Therefore we propose a method to extract the 3D features combined
with the 3D object reconstruction. This method uses the modified
GPU-based visual hull generation algorithm which disables unnecessary
processes, such as the texture calculation to generate three kinds
of 3D projection maps as the 3D feature: a nearest boundary, a farthest
boundary, and a thickness of the object projected on the base-plane. In
the section of experimental results, we present results of proposed
method on eight human postures: T shape, both hands up, right hand
up, left hand up, hands front, stand, sit and bend, and compare the
computational time of the proposed method with that of the previous
methods.
Abstract: Recently, a lot of attention has been devoted to
advanced techniques of system modeling. PNN(polynomial neural
network) is a GMDH-type algorithm (Group Method of Data
Handling) which is one of the useful method for modeling nonlinear
systems but PNN performance depends strongly on the number of
input variables and the order of polynomial which are determined by
trial and error. In this paper, we introduce GPNN (genetic
polynomial neural network) to improve the performance of PNN.
GPNN determines the number of input variables and the order of all
neurons with GA (genetic algorithm). We use GA to search between
all possible values for the number of input variables and the order of
polynomial. GPNN performance is obtained by two nonlinear
systems. the quadratic equation and the time series Dow Jones stock
index are two case studies for obtaining the GPNN performance.
Abstract: This paper presents a system for discovering
association rules from collections of unstructured documents called
EART (Extract Association Rules from Text). The EART system
treats texts only not images or figures. EART discovers association
rules amongst keywords labeling the collection of textual documents.
The main characteristic of EART is that the system integrates XML
technology (to transform unstructured documents into structured
documents) with Information Retrieval scheme (TF-IDF) and Data
Mining technique for association rules extraction. EART depends on
word feature to extract association rules. It consists of four phases:
structure phase, index phase, text mining phase and visualization
phase. Our work depends on the analysis of the keywords in the
extracted association rules through the co-occurrence of the keywords
in one sentence in the original text and the existing of the keywords
in one sentence without co-occurrence. Experiments applied on a
collection of scientific documents selected from MEDLINE that are
related to the outbreak of H5N1 avian influenza virus.