Asrar Institute of Higher Education
Asrar Institute of Higher Education
Abedi A.,Ferdowsi University of Mashhad |
Monsefi R.,Ferdowsi University of Mashhad |
Zadeh D.Z.,Asrar Institute of Higher Education
2016 6th International Conference on Computer and Knowledge Engineering, ICCKE 2016 | Year: 2016
over the last few years, manifold clustering has attracted considerable interest in high-dimensional data clustering. However achieving accurate clustering results that match user desires and data structure is still an open problem. One way to do so is incorporating additional information that indicate relation between data objects. In this paper we propose a method for constrained clustering that take advantage of pairwise constraints. It first solves an optimization program to construct an affinity matrix according to pairwise constraints and manifold structure of data, then applies spectral clustering to find data clusters. Experiments demonstrated that our algorithm outperforms other related algorithms in face image datasets and has comparable results on hand-written digit datasets. © 2016 IEEE.
Zabihzadeh D.,Asrar Institute of Higher Education |
Moattar M.H.,Islamic Azad University at Mashhad
International Journal of Speech Technology | Year: 2014
Speaker verification has been studied widely from different points of view, including accuracy, robustness and being real-time. Recent studies have turned toward better feature stability and robustness. In this paper we study the effect of nonlinear manifold based dimensionality reduction for feature robustness. Manifold learning is a popular recent approach for nonlinear dimensionality reduction. Algorithms for this task are based on the idea that each data point may be described as a function of only a few parameters. Manifold learning algorithms attempt to uncover these parameters in order to find a low-dimensional representation of the data. From the manifold based dimension reduction approaches, we applied the widely used Isometric mapping (Isomap) algorithm. Since in the problem of speaker verification, the input utterance is compared with the model of the claiming client, a speaker dependent feature transformation would be beneficial for deciding on the identity of the speaker. Therefore, our first contribution is to use Isomap dimension reduction approach in the speaker dependent context and compare its performance with two other widely used approaches, namely principle component analysis and factor analysis. The other contribution of our work is to perform the nonlinear transformation in a speaker-dependent framework. We evaluated this approach in a GMM based speaker verification framework using Tfarsdat Telephone speech dataset for different noises and SNRs and the evaluations have shown reliability and robustness even in low SNRs. The results also show better performance for the proposed Isomap approach compared to the other approaches. © 2014 Springer Science+Business Media New York.
Mahdinejad K.,Islamic Azad University at Bojnūrd |
Seghaleh M.Z.,Asrar Institute of Higher education
Life Science Journal | Year: 2013
Time delay estimation (TDE) has been a research topic of significant practical importance in many fields (radar, sonar, seismology, geophysics, ultrasonic, hands-free communications, etc.). It is a first stage that feeds into subsequent processing blocks for identifying, localizing, and tracking radiating sources. This paper presents an implementation of TDE using different weighted generalized cross correlation and make a comparison between them. We will discuss the effect of length of observation interval on accuracy and speed of estimation and the influence of distance of microphones and sound source on estimation error based on some experimental results in room acoustic environments with reverberation and noise.
Reihani E.,Asrar Institute of Higher Education |
Sarikhani A.,Florida International University |
Davodi M.,Asrar Institute of Higher Education |
Davodi M.,Florida International University
International Journal of Electrical Power and Energy Systems | Year: 2012
With the growth of electrical energy demand, providing reliable energy without interruption has become very important nowadays. Maintenance scheduling of generating units is one of the crucial factors in delivering reliable electrical energy to the vital industrial and urban loads. As number of generating units and constraints over their operation is increasing, there is growing need for developing new methods for planning optimal outage of generating units for maintenance. This paper presents a hybrid evolutionary algorithm to tackle the reliability based generator maintenance scheduling problem. Uncertainties in the generating units and the load variations are included so that a more realistic scheduling is obtained. Maintenance scheduling problem is a large scale constrained optimization problem with a large number of variables which needs novel methods to cope with it. A new local search method which is derived from Extremal Optimization (EO) and Genetic Algorithm (GA) is presented to tackle the problem. The proposed method can be used as a local optimizer to further improve the potential solutions in the GA. The proposed method, Hill Climbing Technique (HCT), GA and their hybrid approaches are applied to the IEEE Reliability Test System (RTS) and the obtained results are discussed. © 2012 Elsevier Ltd. All rights reserved.