|
Paper Title |
Authors |
Conference |
Abstract |
|
|
Efficient methods for decreasing training and testing times in speaker verification systems |
M. Raissi Dehkordi
M.M. Homayonpour
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Gaussian Mixture Model (GMM) models feature space using mean vector and covariance matrix of probability
density functions of feature vectors. In this paper, improved Vector Quantization and Covariance Matrix methods were
compared ... more
Gaussian Mixture Model (GMM) models feature space using mean vector and covariance matrix of probability
density functions of feature vectors. In this paper, improved Vector Quantization and Covariance Matrix methods were
compared to GMM. Covariance Matrix model considers shape of probability density functions, while Vector
quantization method models position of probability density functions in feature space. In this paper different methods of
Covariance matrix model including Arithmetic-harmonic sphericity measure and Divergence Shape measure were
examined to evaluate scores in speaker verification task.. Experimental results show that Arithmetic-harmonic
sphericity measure outperforms Divergence shape measure. A novel vector quantization approach was also presented in
this paper. This approach is based on comparing codebook obtained from training data to codebook obtained from test
data. Results show that recent approach has a better performance compared to traditional vector quantization approach.
Also, the results show that Covariance matrix model outperforms improved Vector Quantization and GMM. less
Gaussian Mixture Model (GMM) models feature space using mean vector and covariance matrix of probability
density functions of feature vectors. In this paper, improved Vector Quantization and Covariance Matrix methods were
compared ... more |
خرید مقاله
|
|
Touring a Sequence of Polygons in Weighted Regions |
Amir Hedayaty
Salman Parsa
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Given a subdivision of plane into convex polygon regions, a sequence of polygons to meet,
a start point s, and a target point t, we are interested in determining the shortest ... more
Given a subdivision of plane into convex polygon regions, a sequence of polygons to meet,
a start point s, and a target point t, we are interested in determining the shortest weighted
path on this plane which starts at s, visits each of the polygons in the given order, and ends
at t. The length of a path in weighted regions is de¯ned as the sum of the lengths of the
sub-paths within each region. We will present an approximation algorithm with maximum
± cost additive. Our algorithm is based on the shortest weighted path algorithm proposed
by Mata and Mitchel [2]. The algorithm runs in O(((n3LW +RW) k
± )3) time, where n is the
number of vertices of the region boundaries, L is the longest boundary, W is the maximum
weight in the region, R is the sum of the perimeters of the regions, and k is the number of
polygons. The main idea in the algorithm is to add Steiner points on the region boundaries
and polygon edges. In addition, we will also present a solution to the query version of this
problem. We will extend our result in unweighted version of the \Touring a Sequence of
Polygons" problem [3]. We will give an approximation algorithm to solve the general case of
the problem (with non-convex intersecting polygons). less
Given a subdivision of plane into convex polygon regions, a sequence of polygons to meet,
a start point s, and a target point t, we are interested in determining the shortest ... more |
خرید مقاله
|
|
Dynamic Task Scheduling Algorithm with Minimum Communication Cost and Processor Ready Time in Heterogeneous Systems |
Hadi Shahriar Shahhoseini
Hamidreza Abed Masroorkhah
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Task scheduling is an NP-complete problem. A large number of heuristic approaches for finding the best
result have been presented in the literatures. Most of the algorithms are proposed for the ... more
Task scheduling is an NP-complete problem. A large number of heuristic approaches for finding the best
result have been presented in the literatures. Most of the algorithms are proposed for the homogeneous
systems and a few ones can be applied for the heterogeneous systems. In this paper, a synchronous task
scheduling for heterogeneous systems called Dynamic Task Scheduling with Minimum Communication Cost
and Processor Ready Time (DMCR) is proposed. DMCR can obtain exact values of communication cost and
processors ready times by task selection synchronously with machine selection. Using exact values instead of
conjectural values that used in the other scheduling approaches provides better results in heterogeneous
systems. Simulation shows 5-10% improvement comparing HEFT and 12-18% improvement comparing
HCPT. less
Task scheduling is an NP-complete problem. A large number of heuristic approaches for finding the best
result have been presented in the literatures. Most of the algorithms are proposed for the ... more |
خرید مقاله
|
|
Development of a Computerized System to Detect Patterns of Interstitial Lung Diseases in HRCT Images |
Rahil Garnavi
Hamid Abrishami Moghaddam
Masoomeh Giti
Ahmad Baraani
Samad Ghodrati
Enayat Safavi
Majid Shakiba
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Interstitial Lung Diseases are a complex grouping of disorders with subtle and complex radiographic findings. The
goal of this research is to develop a computerized system to analyze and evaluate the ... more
Interstitial Lung Diseases are a complex grouping of disorders with subtle and complex radiographic findings. The
goal of this research is to develop a computerized system to analyze and evaluate the lung HRCT images, using image
processing and texture analysis techniques. We aim to recognize and quantitatively analyze the three abnormal lung
patterns; “honey-comb”, “ground-glass” and “reticular”. Main steps of the research include lung image segmentation,
feature extraction, pattern classification and finally, finding the correlation between PFT parameters and computerized
data. The system was tested on 160 HRCT scans comprising eight patients with different severities in ILD. The system
shows a partial correlation between computerized data and PFT findings. less
Interstitial Lung Diseases are a complex grouping of disorders with subtle and complex radiographic findings. The
goal of this research is to develop a computerized system to analyze and evaluate the ... more |
خرید مقاله
|
|
Semi-Human Instinctive Artificial Intelligence (SHIAI) |
Asadollah Norouzi
S. Mohammad S. Mohammad
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Providing robots (or any other intelligent embedded system) with manlike instincts will bring major issues of
today artificial intelligence out of a deadlock. This paper proposes a nondeterministic decision making theory
based ... more
Providing robots (or any other intelligent embedded system) with manlike instincts will bring major issues of
today artificial intelligence out of a deadlock. This paper proposes a nondeterministic decision making theory
based on Semi Human Instincts implemented by learned potential fields, using neural networks and fuzzy logic
offline and online learning algorithms, which enable the agent to perform in anonymous, dynamic and nondeterministic
environments. SHI-AI is like a newly born baby who uses his/her instincts and will gradually
become more and more intelligent as the brain learns more about its environment. The use of a new world
modeling method called ARPL (Agent Relative Polar Localization) in SHI-AI enables the agent to perform
better within anonymous environments where positioning is an important and complex issue. less
Providing robots (or any other intelligent embedded system) with manlike instincts will bring major issues of
today artificial intelligence out of a deadlock. This paper proposes a nondeterministic decision making theory
based ... more |
خرید مقاله
|
|
Combinition Of HMM and Neural Network for Phoneme Boundaries Refinement in Persian Database |
M. M. Homayounpour
B. Bakhtiyari
M. Namnabat
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
In this paper, a post-refining method using MLP neural network is proposed for phonetic segmentation and time
alignment of speech databases. The main issue of the work involves the refinement of ... more
In this paper, a post-refining method using MLP neural network is proposed for phonetic segmentation and time
alignment of speech databases. The main issue of the work involves the refinement of an initial estimation of phoneme
boundaries. Initial phoneme boundaries are provided by a time alignment technique using Hidden Markov Model
(HMM) and Multi-layer perceptron (MLP) is used to refine the initial phone boundaries. In fact, MLP neural network
tries to model information in transitions between phonemes. The optimum partitioning of the entire phonetic transition
space is constructed from the standpoint of minimizing the overall deviation from hand labeled positions. The
experimental results show that the proposed method increases boundary estimation performance and achieves high
accuracy compared to manual segmentation. less
In this paper, a post-refining method using MLP neural network is proposed for phonetic segmentation and time
alignment of speech databases. The main issue of the work involves the refinement of ... more |
خرید مقاله
|
|
3TierP2P: A Flexible and Extendible Framework for P2P Application |
Abolhassan Shamsaie
Jafar Habibi
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Peer to peer networks are overlay networks in which all their peers share some resources to participate in a
cumulative service. Now, there exist so many applications which use these distributed ... more
Peer to peer networks are overlay networks in which all their peers share some resources to participate in a
cumulative service. Now, there exist so many applications which use these distributed resources to provide
distributed service. However there is no standard for these networks, but there exist a lot of increasing interests
in this area. This increasing interest and the lack of standard for these networks were the main motivation of this
paper to design and implement a new framework for P2P applications. We considered some goals for our
framework; consist of simplicity, flexibility, extendibility and maintainability, to achieve an appropriate model.
According to these goals and the main idea of the two famous models called three-tier client-server and sevenlayer
OSI model, we proposed a new model for our framework so called 3TierP2P. This model has three tiers in
which, each tier consists of some layers. less
Peer to peer networks are overlay networks in which all their peers share some resources to participate in a
cumulative service. Now, there exist so many applications which use these distributed ... more |
خرید مقاله
|
|
Robust speaker identification based on filtering and spectral peaks in autocorrelation domain |
Amir Hossein Hadjahmadi
Mohammad Mehdi Homayounpour
Gholamreza Farahani
Seyed Mohammad Ahadi
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Noise robustness in speaker recognition has attracted a great deal of interest. This paper describes a novel technique for
noise robust Speaker Identification. In this technique, initially, the speech autocorrelation sequence ... more
Noise robustness in speaker recognition has attracted a great deal of interest. This paper describes a novel technique for
noise robust Speaker Identification. In this technique, initially, the speech autocorrelation sequence is computed and
then, the effect of noise is suppressed using a high pass filter in autocorrelation domain. Finally, the speech feature set is
found using the spectral peaks of this filtered autocorrelation sequence. These Features are robust for speech recognition
task. In this paper we applied them to speaker identification task and found that this features are more robust than
MFCC features. For example in a test of Farsdat speech database, after 10dB corruption of speech signal using babble
noise, it was observed that these features decrease the error rate for more than 22%. less
Noise robustness in speaker recognition has attracted a great deal of interest. This paper describes a novel technique for
noise robust Speaker Identification. In this technique, initially, the speech autocorrelation sequence ... more |
خرید مقاله
|
|
Letter to Sound Using Classification and Regression Trees |
Mohammad Mehdi Homayounpour
Mohammad Izadi
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
Letter To Sound (LTS) is one of the necessary modules in a modern Text To Speech (TTS) system. In this paper
Classification and Regression Trees (CART) are used for Farsi Letter ... more
Letter To Sound (LTS) is one of the necessary modules in a modern Text To Speech (TTS) system. In this paper
Classification and Regression Trees (CART) are used for Farsi Letter to Sound conversion. LTS CART tree is trained
using parameters extracted from training text. This tree is then used for determining phone sequence of those words
whose phonetic transcription do not exist in TTS lexicon. For evaluation of the LTS technique, different Farsi databases
including words and their phonetic transcriptions are used. A LTS performance of 70% was obtained. A good selection
of training database and an appropriate coverage of Farsi phonetic features are important aspects for obtaining a good
system for Farsi LTS. less
Letter To Sound (LTS) is one of the necessary modules in a modern Text To Speech (TTS) system. In this paper
Classification and Regression Trees (CART) are used for Farsi Letter ... more |
خرید مقاله
|
|
VIDEO DENOISING BASED ON A LOCAL GAUSSIAN DISTRIBUTION IN 3-D COMPLEX WAVELET DOMAIN |
Hossein Rabbani
Mansur Vafadust
|
دوازدهمین کنفرانس بین‌المللی سالانه انجمن کامپیوتر ایران |
In this paper we present a new video denoising algorithm that model distribution of wavelet coefficients in each
subband with a Gaussian probability density function (pdf) that its variance is local ... more
In this paper we present a new video denoising algorithm that model distribution of wavelet coefficients in each
subband with a Gaussian probability density function (pdf) that its variance is local (It means that we use a
separate Gaussian pdf for each pixel of each subband). This pdf is capable of modeling the heavy-tailed nature of
wavelet coefficients and the empirically observed correlation between the coefficient amplitudes. Within this
framework, we describe a novel method for video denoising based on designing a maximum a posteriori (MAP)
estimator, which relies on the zero-mean Gaussian random variables with high local correlation. Because
separate 3-D transforms, such as ordinary 3-D wavelet transforms, have visual artifacts that reduce their
performance in applications, we perform our algorithm in 3-D complex wavelet transform. This non-separable
and oriented transform produces a motion-based multiscale decomposition for video that isolates motion along
different directions in its subbands and prevents from directions mixing that appear in subbands of 3-D ordinary
wavelet transform. In addition, we use our denoising algorithm in 2-D complex wavelet transform, where the 2-D
transform is applied to each frame individually. Although our method is simple in its implementation, our
denoising results achieve better performance than several methods visually and regarding peak signal-to-noise
ratio (PSNR). less
In this paper we present a new video denoising algorithm that model distribution of wavelet coefficients in each
subband with a Gaussian probability density function (pdf) that its variance is local ... more |
خرید مقاله
|
.gif){kind=icon}
