Abstract: In this paper we propose segmentation approach based
on Vector Quantization technique. Here we have used Kekre-s fast
codebook generation algorithm for segmenting low-altitude aerial
image. This is used as a preprocessing step to form segmented
homogeneous regions. Further to merge adjacent regions color
similarity and volume difference criteria is used. Experiments
performed with real aerial images of varied nature demonstrate that
this approach does not result in over segmentation or under
segmentation. The vector quantization seems to give far better results
as compared to conventional on-the-fly watershed algorithm.
Abstract: Interaction Model plays an important role in Modelbased
Intelligent Interface Agent Architecture for developing
Intelligent User Interface. In this paper we are presenting some
improvements in the algorithms for development interaction model of
interface agent including: the action segmentation algorithm, the
action pair selection algorithm, the final action pair selection
algorithm, the interaction graph construction algorithm and the
probability calculation algorithm. The analysis of the algorithms also
presented. At the end of this paper, we introduce an experimental
program called “Personal Transfer System".
Abstract: This paper proposes new enhancement models to the
methods of nonlinear anisotropic diffusion to greatly reduce speckle
and preserve image features in medical ultrasound images. By
incorporating local physical characteristics of the image, in this case
scatterer density, in addition to the gradient, into existing tensorbased
image diffusion methods, we were able to greatly improve the
performance of the existing filtering methods, namely edge
enhancing (EE) and coherence enhancing (CE) diffusion. The new
enhancement methods were tested using various ultrasound images,
including phantom and some clinical images, to determine the
amount of speckle reduction, edge, and coherence enhancements.
Scatterer density weighted nonlinear anisotropic diffusion
(SDWNAD) for ultrasound images consistently outperformed its
traditional tensor-based counterparts that use gradient only to weight
the diffusivity function. SDWNAD is shown to greatly reduce
speckle noise while preserving image features as edges, orientation
coherence, and scatterer density. SDWNAD superior performances
over nonlinear coherent diffusion (NCD), speckle reducing
anisotropic diffusion (SRAD), adaptive weighted median filter
(AWMF), wavelet shrinkage (WS), and wavelet shrinkage with
contrast enhancement (WSCE), make these methods ideal
preprocessing steps for automatic segmentation in ultrasound
imaging.
Abstract: Automatic segmentation of skin lesions is the first step
towards the automated analysis of malignant melanoma. Although
numerous segmentation methods have been developed, few studies
have focused on determining the most effective color space for
melanoma application. This paper proposes an automatic segmentation
algorithm based on color space analysis and clustering-based histogram
thresholding, a process which is able to determine the optimal
color channel for detecting the borders in dermoscopy images. The
algorithm is tested on a set of 30 high resolution dermoscopy images.
A comprehensive evaluation of the results is provided, where borders
manually drawn by four dermatologists, are compared to automated
borders detected by the proposed algorithm, applying three previously
used metrics of accuracy, sensitivity, and specificity and a new metric
of similarity. By performing ROC analysis and ranking the metrics,
it is demonstrated that the best results are obtained with the X and
XoYoR color channels, resulting in an accuracy of approximately
97%. The proposed method is also compared with two state-of-theart
skin lesion segmentation methods.
Abstract: The class of geometric deformable models, so-called
level sets, has brought tremendous impact to medical imagery. In
this paper we present yet another application of level sets to medical
imaging. The method we give here will in a way modify the speed
term in the standard level sets equation of motion. To do so we
build a potential based on the distance and the gradient of the
image we study. In turn the potential gives rise to the force field:
F~F(x, y) = P
∀(p,q)∈I
((x, y) - (p, q)) |ÔêçI(p,q)|
|(x,y)-(p,q)|
2 . The direction
and intensity of the force field at each point will determine the
direction of the contour-s evolution. The images we used to test
our method were produced by the Univesit'e de Sherbrooke-s PET
scanners.
Abstract: Effectiveness of Artificial Neural Networks (ANN)
and Support Vector Machines (SVM) classifiers for fault diagnosis of
rolling element bearings are presented in this paper. The
characteristic features of vibration signals of rotating driveline that
was run in its normal condition and with faults introduced were used
as input to ANN and SVM classifiers. Simple statistical features such
as standard deviation, skewness, kurtosis etc. of the time-domain
vibration signal segments along with peaks of the signal and peak of
power spectral density (PSD) are used as features to input the ANN
and SVM classifier. The effect of preprocessing of the vibration
signal by Discreet Wavelet Transform (DWT) prior to feature
extraction is also studied. It is shown from the experimental results
that the performance of SVM classifier in identification of bearing
condition is better then ANN and pre-processing of vibration signal
by DWT enhances the effectiveness of both ANN and SVM classifier
Abstract: Understanding the number of people and the flow of
the persons is useful for efficient promotion of the institution
managements and company-s sales improvements. This paper
introduces an automated method for counting passerby using virtualvertical
measurement lines. The process of recognizing a passerby is
carried out using an image sequence obtained from the USB camera.
Space-time image is representing the human regions which are
treated using the segmentation process. To handle the problem of
mismatching, different color space are used to perform the template
matching which chose automatically the best matching to determine
passerby direction and speed. A relation between passerby speed and
the human-pixel area is used to distinguish one or two passersby. In
the experiment, the camera is fixed at the entrance door of the hall in
a side viewing position. Finally, experimental results verify the
effectiveness of the presented method by correctly detecting and
successfully counting them in order to direction with accuracy of
97%.
Abstract: Optical character recognition of cursive scripts
presents a number of challenging problems in both segmentation and
recognition processes in different languages, including Persian. In
order to overcome these problems, we use a newly developed Persian
word segmentation method and a recognition-based segmentation
technique to overcome its segmentation problems. This method is
robust as well as flexible. It also increases the system-s tolerances to
font variations. The implementation results of this method on a
comprehensive database show a high degree of accuracy which meets
the requirements for commercial use. Extended with a suitable pre
and post-processing, the method offers a simple and fast framework
to develop a full OCR system.
Abstract: In this paper we have proposed three and two
stage still gray scale image compressor based on BTC. In our
schemes, we have employed a combination of four techniques
to reduce the bit rate. They are quad tree segmentation, bit
plane omission, bit plane coding using 32 visual patterns and
interpolative bit plane coding. The experimental results show
that the proposed schemes achieve an average bit rate of 0.46
bits per pixel (bpp) for standard gray scale images with an
average PSNR value of 30.25, which is better than the results
from the exiting similar methods based on BTC.
Abstract: The colonic tissue is a complicated dynamic system
and the colonic activities it generates are composed of irregular
segmental waves, which are referred to as erratic fluctuations or spikes.
They are also highly irregular with subunit fractal structure. The
traditional time-frequency domain statistics like the averaged
amplitude, the motility index and the power spectrum, etc. are
insufficient to describe such fluctuations. Thus the fractal
box-counting dimension is proposed and the fractal scaling behaviors
of the human colonic pressure activities under the physiological
conditions are studied. It is shown that the dimension of the resting
activity is smaller than that of the normal one, whereas the clipped
version, which corresponds to the activity of the constipation patient,
shows with higher fractal dimension. It may indicate a practical
application to assess the colonic motility, which is often indicated by
the colonic pressure activity.
Abstract: In this paper a new concept of partial complement of a graph G is introduced and using the same a new graph parameter, called completion number of a graph G, denoted by c(G) is defined. Some basic properties of graph parameter, completion number, are studied and upperbounds for completion number of classes of graphs are obtained , the paper includes the characterization also.
Abstract: In this paper three different approaches for person
verification and identification, i.e. by means of fingerprints, face and
voice recognition, are studied. Face recognition uses parts-based
representation methods and a manifold learning approach. The
assessment criterion is recognition accuracy. The techniques under
investigation are: a) Local Non-negative Matrix Factorization
(LNMF); b) Independent Components Analysis (ICA); c) NMF with
sparse constraints (NMFsc); d) Locality Preserving Projections
(Laplacianfaces). Fingerprint detection was approached by classical
minutiae (small graphical patterns) matching through image
segmentation by using a structural approach and a neural network as
decision block. As to voice / speaker recognition, melodic cepstral
and delta delta mel cepstral analysis were used as main methods, in
order to construct a supervised speaker-dependent voice recognition
system. The final decision (e.g. “accept-reject" for a verification
task) is taken by using a majority voting technique applied to the
three biometrics. The preliminary results, obtained for medium
databases of fingerprints, faces and voice recordings, indicate the
feasibility of our study and an overall recognition precision (about
92%) permitting the utilization of our system for a future complex
biometric card.
Abstract: We report in this paper the model adopted by our
system of continuous speech recognition in Arab language SySRA
and the results obtained until now. This system uses the database
Arabdic-10 which is a corpus of word for the Arab language and
which was manually segmented. Phonetic decoding is represented
by an expert system where the knowledge base is translated in the
form of production rules. This expert system transforms a vocal
signal into a phonetic lattice. The higher level of the system takes
care of the recognition of the lattice thus obtained by deferring it in
the form of written sentences (orthographical Form). This level
contains initially the lexical analyzer which is not other than the
module of recognition. We subjected this analyzer to a set of
spectrograms obtained by dictating a score of sentences in Arab
language. The rate of recognition of these sentences is about 70%
which is, to our knowledge, the best result for the recognition of the
Arab language. The test set consists of twenty sentences from four
speakers not having taken part in the training.
Abstract: A passive system "Qanat" is collection of some
underground wells. A mother-well was dug in a place far from the
city where they could reach to the water table maybe 100 meters
underground, they dug other wells to direct water toward the city,
with minimum possible gradient. Using the slope of the earth they
could bring water close to the surface in the city. The source of water
or the appearance of Qanat, land slope and the ownership lines are
the important and effective factors in the formation of routes and the
segment division of lands to the extent that making use of Qanat as
the techniques of extracting underground waters creates a channel of
routes with an organic order and hierarchy coinciding the slope of
land and it also guides the Qanat waters in the tradition texture of salt
desert and border provinces of it. Qanats are excavated in a specified
distinction from each other. The quantity of water provided by
Qanats depends on the kind of land, distance from mountain,
geographical situation of them and the rate of water supply from the
underground land. The rate of underground waters, possibility of
Qanat excavation, number of Qanats and rate of their water supply
from one hand and the quantity of cultivable fertile lands from the
other hand are the important natural factors making the size of cities.
In the same manner the cities with several Qanats have multi central
textures. The location of cities is in direct relation with land quality,
soil fertility and possibility of using underground water by excavating
Qanats. Observing the allowable distance for Qanat watering is a
determining factor for distance between villages and cities.
Topography, land slope, soil quality, watering system, ownership,
kind of cultivation, etc. are the effective factors in directing Qanats
for excavation and guiding water toward the cultivable lands and it
also causes the formation of different textures in land division of
farming provinces. Several divisions such as orderly and wide, inorderly,
thin and long, comb like, etc. are the introduction to organic
order. And at the same time they are complete coincidence with
environmental conditions in the typical development of ecological
architecture and planning in the traditional cities and settlements
order.
Abstract: This study aims to segment objects using the K-means
algorithm for texture features. Firstly, the algorithm transforms color
images into gray images. This paper describes a novel technique for
the extraction of texture features in an image. Then, in a group of
similar features, objects and backgrounds are differentiated by using
the K-means algorithm. Finally, this paper proposes a new object
segmentation algorithm using the morphological technique. The
experiments described include the segmentation of single and multiple
objects featured in this paper. The region of an object can be
accurately segmented out. The results can help to perform image
retrieval and analyze features of an object, as are shown in this paper.
Abstract: The compression-absorption heat pump (C-A HP), one
of the promising heat recovery equipments that make process hot
water using low temperature heat of wastewater, was evaluated by
computer simulation. A simulation program was developed based on
the continuity and the first and second laws of thermodynamics. Both
the absorber and desorber were modeled using UA-LMTD method. In
order to prevent an unfeasible temperature profile and to reduce
calculation errors from the curved temperature profile of a mixture,
heat loads were divided into lots of segments. A single-stage
compressor was considered. A compressor cooling load was also
taken into account. An isentropic efficiency was computed from the
map data. Simulation conditions were given based on the system
consisting of ordinarily designed components. The simulation results
show that most of the total entropy generation occurs during the
compression and cooling process, thus suggesting the possibility that
system performance can be enhanced if a rectifier is introduced.
Abstract: The increasing demand for sufficient and clean
energy forces industrial and service companies to align their strategies towards efficient consumption. This trend refers also to the
residential building sector. There, large amounts of energy consumption are caused by house and facility heating. Many of the
operated hot water heating systems lack hydraulic balanced working
conditions for heat distribution and –transmission and lead to
inefficient heating. Through hydraulic balancing of heating systems,
significant energy savings for primary and secondary energy can be
achieved. This paper addresses the use of KNX-technology (Smart
Buildings) in residential buildings to ensure a dynamic adaption of
hydraulic system's performance, in order to increase the heating
system's efficiency. In this paper, the procedure of heating system
segmentation into hydraulically independent units (meshes) is
presented. Within these meshes, the heating valve are addressed and
controlled by a central facility server. Feasibility criteria towards
such drivers will be named. The dynamic hydraulic balance is
achieved by positioning these valves according to heating loads, that
are generated from the temperature settings in the corresponding
rooms. The energetic advantages of single room heating control
procedures, based on the application FacilityManager, is presented.
Abstract: This paper addresses the development of an intelligent vision system for human-robot interaction. The two novel contributions of this paper are 1) Detection of human faces and 2) Localizing the eye. The method is based on visual attributes of human skin colors and geometrical analysis of face skeleton. This paper introduces a spatial domain filtering method named ?Fuzzily skewed filter' which incorporates Fuzzy rules for deciding the gray level of pixels in the image in their neighborhoods and takes advantages of both the median and averaging filters. The effectiveness of the method has been justified over implementing the eye tracking commands to an entertainment robot, named ''AIBO''.
Abstract: Electrocardiogram (ECG) segmentation is necessary
to help reduce the time consuming task of manually annotating
ECG-s. Several algorithms have been developed to segment the ECG
automatically. We first review several of such methods, and then
present a new single lead segmentation method based on Adaptive
piecewise constant approximation (APCA) and Piecewise derivative
dynamic time warping (PDDTW). The results are tested on the QT
database. We compared our results to Laguna-s two lead method. Our
proposed approach has a comparable mean error, but yields a slightly
higher standard deviation than Laguna-s method.
Abstract: In this paper we introduce a novel kernel classifier
based on a iterative shrinkage algorithm developed for compressive
sensing. We have adopted Bregman iteration with soft and hard
shrinkage functions and generalized hinge loss for solving l1 norm
minimization problem for classification. Our experimental results
with face recognition and digit classification using SVM as the
benchmark have shown that our method has a close error rate
compared to SVM but do not perform better than SVM. We have
found that the soft shrinkage method give more accuracy and in some
situations more sparseness than hard shrinkage methods.