Abstract: Clusters of microcalcifications in mammograms are an
important sign of breast cancer. This paper presents a complete
Computer Aided Detection (CAD) scheme for automatic detection of
clustered microcalcifications in digital mammograms. The proposed
system, MammoScan μCaD, consists of three main steps. Firstly
all potential microcalcifications are detected using a a method for
feature extraction, VarMet, and adaptive thresholding. This will also
give a number of false detections. The goal of the second step,
Classifier level 1, is to remove everything but microcalcifications.
The last step, Classifier level 2, uses learned dictionaries and sparse
representations as a texture classification technique to distinguish
single, benign microcalcifications from clustered microcalcifications,
in addition to remove some remaining false detections. The system
is trained and tested on true digital data from Stavanger University
Hospital, and the results are evaluated by radiologists. The overall
results are promising, with a sensitivity > 90 % and a low false
detection rate (approx 1 unwanted pr. image, or 0.3 false pr. image).
Abstract: In this paper, an automated system is presented for
identification and separation of plastic resins based on near infrared
(NIR) reflectance spectroscopy. For identification and separation
among resins, a "Two-Filter" identification method is proposed that
is capable to distinguish among polyethylene terephthalate (PET),
high density polyethylene (HDPE), polyvinyl chloride (PVC),
polypropylene (PP) and polystyrene (PS). Through surveying effects
of parameters such as surface contamination, sample thickness, label
and cap existence, it was obvious that the "Two-Filter" method has a
high efficiency in identification of resins. It is shown that accurate
identification and separation of five major resins can be obtained
through calculating the relative reflectance at two wavelengths in the
NIR region.
Abstract: This paper proposes to use ETM+ multispectral data
and panchromatic band as well as texture features derived from the
panchromatic band for land cover classification. Four texture features
including one 'internal texture' and three GLCM based textures
namely correlation, entropy, and inverse different moment were used
in combination with ETM+ multispectral data. Two data sets
involving combination of multispectral, panchromatic band and its
texture were used and results were compared with those obtained by
using multispectral data alone. A decision tree classifier with and
without boosting were used to classify different datasets. Results
from this study suggest that the dataset consisting of panchromatic
band, four of its texture features and multispectral data was able to
increase the classification accuracy by about 2%. In comparison, a
boosted decision tree was able to increase the classification accuracy
by about 3% with the same dataset.
Abstract: The evaluation of conversational agents or chatterbots question answering systems is a major research area that needs much attention. Before the rise of domain-oriented conversational agents based on natural language understanding and reasoning, evaluation is never a problem as information retrieval-based metrics are readily available for use. However, when chatterbots began to become more domain specific, evaluation becomes a real issue. This is especially true when understanding and reasoning is required to cater for a wider variety of questions and at the same time to achieve high quality responses. This paper discusses the inappropriateness of the existing measures for response quality evaluation and the call for new standard measures and related considerations are brought forward. As a short-term solution for evaluating response quality of conversational agents, and to demonstrate the challenges in evaluating systems of different nature, this research proposes a blackbox approach using observation, classification scheme and a scoring mechanism to assess and rank three example systems, AnswerBus, START and AINI.
Abstract: This paper investigates how the use of machine learning techniques can significantly predict the three major dimensions of learner-s emotions (pleasure, arousal and dominance) from brainwaves. This study has adopted an experimentation in which participants were exposed to a set of pictures from the International Affective Picture System (IAPS) while their electrical brain activity was recorded with an electroencephalogram (EEG). The pictures were already rated in a previous study via the affective rating system Self-Assessment Manikin (SAM) to assess the three dimensions of pleasure, arousal, and dominance. For each picture, we took the mean of these values for all subjects used in this previous study and associated them to the recorded brainwaves of the participants in our study. Correlation and regression analyses confirmed the hypothesis that brainwave measures could significantly predict emotional dimensions. This can be very useful in the case of impassive, taciturn or disabled learners. Standard classification techniques were used to assess the reliability of the automatic detection of learners- three major dimensions from the brainwaves. We discuss the results and the pertinence of such a method to assess learner-s emotions and integrate it into a brainwavesensing Intelligent Tutoring System.
Abstract: In this study we focus on improvement performance
of a cue based Motor Imagery Brain Computer Interface (BCI). For
this purpose, data fusion approach is used on results of different
classifiers to make the best decision. At first step Distinction
Sensitive Learning Vector Quantization method is used as a feature
selection method to determine most informative frequencies in
recorded signals and its performance is evaluated by frequency
search method. Then informative features are extracted by packet
wavelet transform. In next step 5 different types of classification
methods are applied. The methodologies are tested on BCI
Competition II dataset III, the best obtained accuracy is 85% and the
best kappa value is 0.8. At final step ordered weighted averaging
(OWA) method is used to provide a proper aggregation classifiers
outputs. Using OWA enhanced system accuracy to 95% and kappa
value to 0.9. Applying OWA just uses 50 milliseconds for
performing calculation.
Abstract: This article presents the development of a neural
network cognitive model for the classification and detection of
different frequency signals. The basic structure of the implemented
neural network was inspired on the perception process that humans
generally make in order to visually distinguish between high and low
frequency signals. It is based on the dynamic neural network concept,
with delays. A special two-layer feedforward neural net structure was
successfully implemented, trained and validated, to achieve
minimum target error. Training confirmed that this neural net
structure descents and converges to a human perception classification
solution, even when far away from the target.
Abstract: In this study, a high accuracy protein-protein interaction
prediction method is developed. The importance of the proposed
method is that it only uses sequence information of proteins while
predicting interaction. The method extracts phylogenetic profiles of
proteins by using their sequence information. Combining the phylogenetic
profiles of two proteins by checking existence of homologs
in different species and fitting this combined profile into a statistical
model, it is possible to make predictions about the interaction status
of two proteins.
For this purpose, we apply a collection of pattern recognition
techniques on the dataset of combined phylogenetic profiles of protein
pairs. Support Vector Machines, Feature Extraction using ReliefF,
Naive Bayes Classification, K-Nearest Neighborhood Classification,
Decision Trees, and Random Forest Classification are the methods
we applied for finding the classification method that best predicts
the interaction status of protein pairs. Random Forest Classification
outperformed all other methods with a prediction accuracy of 76.93%
Abstract: This paper presents a technique for diagnosis of the abdominal aorta aneurysm in magnetic resonance imaging (MRI) images. First, our technique is designed to segment the aorta image in MRI images. This is a required step to determine the volume of aorta image which is the important step for diagnosis of the abdominal aorta aneurysm. Our proposed technique can detect the volume of aorta in MRI images using a new external energy for snakes model. The new external energy for snakes model is calculated from Law-s texture. The new external energy can increase the capture range of snakes model efficiently more than the old external energy of snakes models. Second, our technique is designed to diagnose the abdominal aorta aneurysm by Bayesian classifier which is classification models based on statistical theory. The feature for data classification of abdominal aorta aneurysm was derived from the contour of aorta images which was a result from segmenting of our snakes model, i.e., area, perimeter and compactness. We also compare the proposed technique with the traditional snakes model. In our experiment results, 30 images are trained, 20 images are tested and compared with expert opinion. The experimental results show that our technique is able to provide more accurate results than 95%.
Abstract: Random and natural textures classification is still
one of the biggest challenges in the field of image processing and
pattern recognition. In this paper, texture feature extraction using
Slant Hadamard Transform was studied and compared to other
signal processing-based texture classification schemes. A
parametric SHT was also introduced and employed for natural
textures feature extraction. We showed that a subtly modified
parametric SHT can outperform ordinary Walsh-Hadamard
transform and discrete cosine transform. Experiments were carried
out on a subset of Vistex random natural texture images using a
kNN classifier.
Abstract: Near-infrared (NIR) spectroscopy is a widely used
method for material identification for laboratory and industrial applications.
While standard spectrometers only allow measurements at
one sampling point at a time, NIR Spectral Imaging techniques can
measure, in real-time, both the size and shape of an object as well as
identify the material the object is made of. The online classification
and sorting of recovered paper with NIR Spectral Imaging (SI)
is used with success in the paper recycling industry throughout
Europe. Recently, the globalisation of the recycling material streams
caused that water-based flexographic-printed newspapers mainly from
UK and Italy appear also in central Europe. These flexo-printed
newspapers are not sufficiently de-inkable with the standard de-inking
process originally developed for offset-printed paper. This de-inking
process removes the ink from recovered paper and is the fundamental
processing step to produce high-quality paper from recovered paper.
Thus, the flexo-printed newspapers are a growing problem for the
recycling industry as they reduce the quality of the produced paper
if their amount exceeds a certain limit within the recovered paper
material.
This paper presents the results of a research project for the
development of an automated entry inspection system for recovered
paper that was jointly conducted by CTR AG (Austria) and PTS
Papiertechnische Stiftung (Germany). Within the project an NIR
SI prototype for the identification of flexo-printed newspaper has
been developed. The prototype can identify and sort out flexoprinted
newspapers in real-time and achieves a detection accuracy
for flexo-printed newspaper of over 95%. NIR SI, the technology the
prototype is based on, allows the development of inspection systems
for incoming goods in a paper production facility as well as industrial
sorting systems for recovered paper in the recycling industry in the
near future.
Abstract: Young patients suffering from Cerebral Palsy are
facing difficult choices concerning heavy surgeries. Diagnosis settled
by surgeons can be complex and on the other hand decision for
patient about getting or not such a surgery involves important
reflection effort. Proposed software combining prediction for
surgeries and post surgery kinematic values, and from 3D model
representing the patient is an innovative tool helpful for both patients
and medicine professionals. Beginning with analysis and
classification of kinematics values from Data Base extracted from
gait analysis in 3 separated clusters, it is possible to determine close
similarity between patients. Prediction surgery best adapted to
improve a patient gait is then determined by operating a suitable
preconditioned neural network. Finally, patient 3D modeling based
on kinematic values analysis, is animated thanks to post surgery
kinematic vectors characterizing the closest patient selected from
patients clustering.
Abstract: As the network based technologies become
omnipresent, demands to secure networks/systems against threat
increase. One of the effective ways to achieve higher security is
through the use of intrusion detection systems (IDS), which are a
software tool to detect anomalous in the computer or network. In this
paper, an IDS has been developed using an improved machine
learning based algorithm, Locally Linear Neuro Fuzzy Model
(LLNF) for classification whereas this model is originally used for
system identification. A key technical challenge in IDS and LLNF
learning is the curse of high dimensionality. Therefore a feature
selection phase is proposed which is applicable to any IDS. While
investigating the use of three feature selection algorithms, in this
model, it is shown that adding feature selection phase reduces
computational complexity of our model. Feature selection algorithms
require the use of a feature goodness measure. The use of both a
linear and a non-linear measure - linear correlation coefficient and
mutual information- is investigated respectively
Abstract: Linear Discrimination Analysis (LDA) is a linear
solution for classification of two classes. In this paper, we propose a
variant LDA method for multi-class problem which redefines the
between class and within class scatter matrices by incorporating a
weight function into each of them. The aim is to separate classes as
much as possible in a situation that one class is well separated from
other classes, incidentally, that class must have a little influence on
classification. It has been suggested to alleviate influence of classes
that are well separated by adding a weight into between class scatter
matrix and within class scatter matrix. To obtain a simple and
effective weight function, ordinary LDA between every two classes
has been used in order to find Fisher discrimination value and passed
it as an input into two weight functions and redefined between class
and within class scatter matrices. Experimental results showed that
our new LDA method improved classification rate, on glass, iris and
wine datasets, in comparison to different versions of LDA.
Abstract: Data mining techniques have been used in medical
research for many years and have been known to be effective. In order
to solve such problems as long-waiting time, congestion, and delayed
patient care, faced by emergency departments, this study concentrates
on building a hybrid methodology, combining data mining techniques
such as association rules and classification trees. The methodology is
applied to real-world emergency data collected from a hospital and is
evaluated by comparing with other techniques. The methodology is
expected to help physicians to make a faster and more accurate
classification of chest pain diseases.
Abstract: Logic based methods for learning from structured data
is limited w.r.t. handling large search spaces, preventing large-sized
substructures from being considered by the resulting classifiers. A
novel approach to learning from structured data is introduced that
employs a structure transformation method, called finger printing, for
addressing these limitations. The method, which generates features
corresponding to arbitrarily complex substructures, is implemented in
a system, called DIFFER. The method is demonstrated to perform
comparably to an existing state-of-art method on some benchmark
data sets without requiring restrictions on the search space.
Furthermore, learning from the union of features generated by finger
printing and the previous method outperforms learning from each
individual set of features on all benchmark data sets, demonstrating
the benefit of developing complementary, rather than competing,
methods for structure classification.
Abstract: Classification of electroencephalogram (EEG) signals
extracted during mental tasks is a technique that is actively pursued
for Brain Computer Interfaces (BCI) designs. In this paper, we
compared the classification performances of univariateautoregressive
(AR) and multivariate autoregressive (MAR) models
for representing EEG signals that were extracted during different
mental tasks. Multilayer Perceptron (MLP) neural network (NN)
trained by the backpropagation (BP) algorithm was used to classify
these features into the different categories representing the mental
tasks. Classification performances were also compared across
different mental task combinations and 2 sets of hidden units (HU): 2
to 10 HU in steps of 2 and 20 to 100 HU in steps of 20. Five different
mental tasks from 4 subjects were used in the experimental study and
combinations of 2 different mental tasks were studied for each
subject. Three different feature extraction methods with 6th order
were used to extract features from these EEG signals: AR
coefficients computed with Burg-s algorithm (ARBG), AR
coefficients computed with stepwise least square algorithm (ARLS)
and MAR coefficients computed with stepwise least square
algorithm. The best results were obtained with 20 to 100 HU using
ARBG. It is concluded that i) it is important to choose the suitable
mental tasks for different individuals for a successful BCI design, ii)
higher HU are more suitable and iii) ARBG is the most suitable
feature extraction method.
Abstract: In this paper, we propose a new method to distinguish
between arousal and relaxation states by using multiple features
acquired from a photoplethysmogram (PPG) and support vector
machine (SVM). To induce arousal and relaxation states in subjects, 2
kinds of sound stimuli are used, and their corresponding biosignals are
obtained using the PPG sensor. Two features–pulse to pulse interval
(PPI) and pulse amplitude (PA)–are extracted from acquired PPG
data, and a nonlinear classification between arousal and relaxation is
performed using SVM.
This methodology has several advantages when compared with
previous similar studies. Firstly, we extracted 2 separate features from
PPG, i.e., PPI and PA. Secondly, in order to improve the classification
accuracy, SVM-based nonlinear classification was performed.
Thirdly, to solve classification problems caused by generalized
features of whole subjects, we defined each threshold according to
individual features.
Experimental results showed that the average classification
accuracy was 74.67%. Also, the proposed method showed the better
identification performance than the single feature based methods.
From this result, we confirmed that arousal and relaxation can be
classified using SVM and PPG features.
Abstract: Power System Security is a major concern in real time
operation. Conventional method of security evaluation consists of
performing continuous load flow and transient stability studies by
simulation program. This is highly time consuming and infeasible
for on-line application. Pattern Recognition (PR) is a promising
tool for on-line security evaluation. This paper proposes a Support
Vector Machine (SVM) based binary classification for static and
transient security evaluation. The proposed SVM based PR approach
is implemented on New England 39 Bus and IEEE 57 Bus systems.
The simulation results of SVM classifier is compared with the other
classifier algorithms like Method of Least Squares (MLS), Multi-
Layer Perceptron (MLP) and Linear Discriminant Analysis (LDA)
classifiers.
Abstract: We present an Electronic Nose (ENose), which is
aimed at identifying the presence of one out of two gases, possibly
detecting the presence of a mixture of the two. Estimation of the
concentrations of the components is also performed for a volatile
organic compound (VOC) constituted by methanol and acetone, for
the ranges 40-400 and 22-220 ppm (parts-per-million), respectively.
Our system contains 8 sensors, 5 of them being gas sensors (of the
class TGS from FIGARO USA, INC., whose sensing element is a tin
dioxide (SnO2) semiconductor), the remaining being a temperature
sensor (LM35 from National Semiconductor Corporation), a
humidity sensor (HIH–3610 from Honeywell), and a pressure sensor
(XFAM from Fujikura Ltd.).
Our integrated hardware–software system uses some machine
learning principles and least square regression principle to identify at
first a new gas sample, or a mixture, and then to estimate the
concentrations. In particular we adopt a training model using the
Support Vector Machine (SVM) approach with linear kernel to teach
the system how discriminate among different gases. Then we apply
another training model using the least square regression, to predict
the concentrations.
The experimental results demonstrate that the proposed
multiclassification and regression scheme is effective in the
identification of the tested VOCs of methanol and acetone with
96.61% correctness. The concentration prediction is obtained with
0.979 and 0.964 correlation coefficient for the predicted versus real
concentrations of methanol and acetone, respectively.