Sharif Digital Repository

This thesis proposes a novel method for enhancing the speech signal based on compressed sensing. Compressed sensing, as a new rapidly growing research field, promises to effectively recover a sparse signal at the rate of below Nyquist rate. This revolutionary technology strongly relies on the sparsity of the signal and incoherency between sensing basis and representation basis. Exact recovery of a sparse signal will be occurred in a situation that the signal of interest sensed randomly and the measurements are also taken based on sparsity level and log factor of the signal dimension.
In this research, compressed sensing method is proposed to reconstruct speech signal and for noise... 

Biometrics has been long known to recognize persons based on their physical and behavioral characteristics. Face recognition (FR) is one of such biometrics that has received a considerable attention in recent years both from the industry and research communities. As the boosting framework has shown good performance in face recognition, it has been adopted in this work. This thesis deals with pattern recognition methods such as linear discriminant analysis (LDA) and machine learning approaches such as boosting which are integrated to overcome the technical limitation of existing FR methods. However, LDA-based methods often suffer from the so-called “small-sample-size” (SSS) problem arising... 

This thesis proposes a novel method for enhancing the image signal based on compressed sensing. Compressed sensing, as a new rapidly growing research field, promises to effectively recover a sparse signal at the rate of below Nyquist rate. This revolutionary technology strongly relies on the sparsity of the signal and incoherency between sensing basis and representation basis. Exact recovery of a sparse signal will be occurred in a situation that the signal of interest sensed randomly and the measurements are also taken based on sparsity level and log factor of the signal dimension. In this research, compressed sensing method is proposed to reduce the noise and reconstruct the image signal.... 

The aim of this study was to investigate the rate of increase in temperature of a phantom equivalent to body tissue by different groups of magnetic iron nanoparticles in the external magnetic field to kill cancer cells based on the hyperthermia method. To achieve this goal, three groups of dextran magnetic nanoparticles with different properties and reduced iron oxide-graphene oxide magnetic nanoparticles by M-rGO supercritical synthesis method were used. After XRD, FTIR, SEM, FESEM, VSM, TEM characterization tests, these materials were placed in a phantom made of agarose gel and with the same properties, in a magnetic field with fixed characteristics for all groups and during the process of... 

The importance of optimizing the extraction process of available resources increases each day due to the increasing energy consumption and the lack of energy resources. Oil and gas are one of the most important sources of energy. Although existing oil and gas resources are thought to be sufficient to meet the growing energy demand for the next few decades, given the non-renewable nature of these resources and the growing demand for oil and gas, it will become much harder to meet the future energy demand. Many existing oil fields are now in the process of maturing, and the discovery of large new oil fields is rare. As a result, new technologies must be used in the future to meet this demand,... 

Magnetic resonance imaging (MRI) is a highly efficient method that can provide acceptable contrast between soft tissues. But the big disadvantage of this method is that its acquisition is slow. To find the reason of being time-consuming should the procedure be surveyed. In the magnetic resonance imaging, location information obtained using phase and frequency encoding gradients. So the output is matrix of image data in the frequency domain, which is called k space. For the formation of the k space phase, need to apply gradients several times and this is the main reason of dullness of the system. Therefore, in this project using software methods try to reduce the scan time as possible. Among... 

Localization of a neutron point source using a designed computer program namely “MCMC-MATURE” is performed. The computer program analyses several detector responses in some certain media by Markov Chain Monte Carlo (MCMC) method and a new iteration algorithm. Identification of the possible regions of source position would be found by analyzing the initial fluxes generated by mesh tally of MCNPX computer code. The designed computer program is capable to generate the flux between detectors. “Regular-Sequential”, “Irregular-Sequential” and “Non-Sequential” are three methods used for sampling the generated random number in two dimensions. Each sample multiplied by a sampling function and lead to... 

Ahwaz is typical of a polluted city. It ranks in the third place after Tehran and Esfahan and with the onslaught Of the dust storms it has acquired the dubious distinction of 1st polluted city in the world. The pollution status of this city was investigated from May 2012 to May 2013. 110 air filters were collected during this period. The air filters were analysed with AAS and INAA techniques. with the use of AAS we measured the concentration of Cd, Cu, Fe, Mg, Mn, Ni and Pb elements. For the use of INAA method on the air filters we used two reactors of MNSR at Esfahan and the 5MW research reactor of Tehran. We managed to measure twenty five elements of Al , As , Br , Ca , Ce , Cl , Co , Cr ,... 

This study considers an online multi-period assortment optimization problem over multiple replenishment cycles. The retailer chooses a subset from N substitutable products and decides how many of each product to order and sell at each time period. Retailer is constrained by a total inventory capacity, a cardinality constraint on the product variety (display space), and predetermined replenishment time intervals. The assortment selection is modeled as a Multi-armed Bandit problem and the customers' choice is modeled by the Multinomial Logit (MNL) choice model with dynamic substitution. The objective is to optimize the revenue by actively learning the censored demand and improving the offering... 

Iranian Household Expenditures and Income Survay(IHEIS) in urban area shows that, average of real hourly wage against the schooling years is declining in the right tail of wage distribution between 1390 and 1384(2005 and 2011). A tentative justification for this observation is that the private rate of return to education at upper quantiles and higher years of schooling have been reduced in the mentioned period. Also, It is observed that the real wage inequality between upper and lower deciles in 1390 is less than 1384. This study intend to investigate the changes in wage inequality between different deciles by different levels of education and rates of return to education in the time... 

Over employment is a common phenomenon in Iran. Over 61% of the labor force with more than 31 years of work experience have worked more than 42 hours a week in year 2000. This amount, increases for the years 2001, 2003 & 2004 as 65.9%, 68.5% & 72.1%. In this research, we examined the possibility of job opportunity reallocation by increasing the overtime premium between over employed and unemployed labor force in Iranian labor market. For this goal, we estimated the determinants of overtime incidence and hours according to single selection & double selection approach, Tunali (1986). Pooled data from urban Iranian labor force- year 2005- was collected and analyzed. Our results clearly show... 

In examining protein-protein interaction networks, we often encounter similar and repetitive schemes. Examination of these designs, which often appear in the form of motifs and similar patterns, reveals important information such as the type of protein linkage and many of the internal similarities between these networks. The ability to recognize these similarities plays an important role in identifying the function of genes, recognizing the relationships between diseases, and making drugs. We know that exact algorithms for examining subgraph isomorphism are np-hard and time-consuming and infeasible in large networks; Therefore, in practice, approximate and heuristic algorithms are used and... 

Deep learning methods have become more popular in the past years. These methods use complex network architectures to model rich, hierarchical datasets. Although most of the research has been centered around Discriminative models, however, recently a lot of research is focused on Deep Generative Models. Two of the pioneering models in this field are Generative Adversarial Networks and Variational Auto-Encoders. In addition, knowing the structure of data helps models to search in a narrower hypothesis space. Most of the structure in datasets are models using Probabilistic Graphical Models. Using this structural information, one can achieve better parameter estimations. In the case of... 

Identifying central node is an important issue in network structural analysis. Nodes with high centrality have a significant impact on spreading of influence and ideas in social networks, activity of nodes in mobile phone networks, and also act as bottlenecks in communication networks. Therefore, identifying k-highest central nodes will be of great interest in many applications. To this end, many exact and approximation algorithms have been recently proposed. The major drawback of these algorithms is that they are not efficient with respect to the tremendous size of today’s networks. Moreover, most of these algorithms assume full knowledge of the network topological structure which is not... 

In many network which relationships between nodes are defined based on the similarity of attributes (Such as the World Wide Web and social networks), extracting information about networks object’s attributes may be difficult or even in many cases impossible.In these cases, predicting unknown attributes based on other objects attributes according to network structure can be extremely useful.Even more, finding a representative sample set of objects and trying to obtain their attributes in order to predict other object’s attributes with this obtained data can be an interesting problem. Finding such a set of objects with minimum size while giving maximizing accuracy in predicting other object’s... 

Many studies has been focused on extracting structural and behavioral properties of complex networks in recent decade. Online Social Networks (OSNs) are one example of complex networks. Nowadays with rapid growth of OSNs such as Facebook and Twitter, the study of OSNs has become an interesting research area. Many of recent OSN studies studied friendship networks. Friendship network is a binary unweighted network, and all of its links have the same importance. But, in reality not only all friendship links are not representative of social interactions, but also the social links have a variety of intimacy, intensity, and etc. So, all links should not be considered equal. Recently researchers... 

Many real-world communication systems such as Internet, online social networks, and brain networks can be modeled as a complex network of interacting dynamical nodes. These networks have non-trivial topological features, i.e., features that do not occur in simple networks such as lattices or random networks. The tremendous growth of Internet and its applications in recent years has resulted in creation of large-scale complex networks involving tens or hundreds of millions of nodes and links. Thus, it may be impossible or costly to obtain a complete picture of these large networks, and sampling methods are essential for practical estimation of network properties. Therefore, in this thesis, we... 

This study proposes a probabilistic approach to prioritize a collection of buildings in a region for seismic retrofit based on regional risk analysis. Risk in this context is the exceedance probability of total economic and social losses, which include both direct and indirect consequences resulting from building damage. The prioritization criterion is determined by the benefit-to-cost ratio of retrofit, where the benefit is measured by the reduction of seismic risk achieved through retrofitting. Hence, the buildings are analyzed both in their current state and after retrofit to compute the risk measure, i.e., the mean total loss. Buildings that yield the highest reduction of the risk... 

This research proposes a probabilistic framework for seismic risk analysis of industrial plants of oil infrastructure using reliability methods. The proposed approach integrates multiple probabilistic models and system reliability to quantify the seismic risk of the process plants. A chain of probabilistic seismic hazard and risk models is utilized for the risk analysis. The Monte Carlo sampling method is used to propagate the significant uncertainty in the hazard event, damage to components and its consequences, and the potential losses incurred by process plant. In each sample of the analysis, hazard models simulate the occurrence, magnitude, and rupture area and location of earthquake... 

The fundamental objective of this study is to present seismic hazard, risk and resilience assessment of a virtual city using sampling and agent-based simulation in two levels of refinement. In this context, hazard is the exceedance probability of ground shaking intensity, risk is exceedance probability of measures such as loss, and resilience is the ability to quickly recover after a hazard event. This study utilizes the Rtx risk and resilience assessment framework established at the Center for Infrastructure Sustainability and Resilience Research (INSURER) at Sharif University of Technology. This framework employs multiple interacting probabilistic models to quantify the risk and...