Sharif Digital Repository

The objective of OCR systems is to retrieve machine-encoded text from a raster image. Despite the abundance of powerful OCR algorithms for English, there are not many for Farsi. Our proposed algorithm is comprised of pre-processing, line detection, sub-word detection and segmentation, feature extraction and classification. Furthermore, hardware implementation and acceleration of this system on a GPGPU is presented. This algorithm was tested on 5 fonts including Titr, Lotus,Yekan, Koodak and Nazanin and an average accuracy above 90% was achieved  

Many mobile applications running on smartphones, wearable devices, tiny autonomous robots and IoT devices would potentially benefit from the accuracy and scalability of deep CNN-based machine learning algorithms. However,performance and energy consumption limitations make the execution of such computationally intensive algorithms on embedded mobile devices prohibitive.We present a GPU-accelerated engine, dubbed mCNN, for execution of trained deep CNNs on mobile platforms. The proposed solution takes the trained model as input and automatically optimizes its parallel implementation on the target mobile platform for efficient use of hardware resources such as mobile GPU threads and SIMD units.... 

Spiking Neural Networks(SNN) are networks which are consisted of layers of neurons, like other typical artificial neural networks. The main difference between SNN and other neural networks is the type of data transportation among neurons which is done by spikes. Spiking neural networks and their models are considered as the nearest networks and neurons to animals’ nervous systems. In aspects of hardware implementation, the type of data transportation in SNN causes them to be ultra-low power. So, implementation of these networks on chips like FPGA and also usage of SNN in applications with high processing load have startling germination, recently. In this work, we have tried to propose some... 

Many chip manufacturers have recently introduced high-performance deep-learning hardware accelerators. In modern GPUs, programmable tensor cores accelerate the heavy operations involved in deep neural networks. This paper presents a novel solution to re-purpose tensor cores in modern GPUs for high-throughput implementation of turbo decoders. Turbo codes closely approach Shannon’s limit on channel capacity, and are widely used in many state-of-the-art wireless systems including satellite communications and mobile communications. Experimental evaluations show that the proposed solution achieves about 1.2 Gbps throughput, which is higher compared to previous GPU-accelerated solutions  

There have been many works on generative networks and image generation in the past few years, but the problem with this work is that there is no control over the generated images. The goal of disentangled image synthesis is to generate new images with specific detail and have control over the generated images. Image-based virtual try-on aims to synthesize the customer image with an in-shop clothes image to acquire seamless and natural try-on results, which have attracted increasing attention. The main procedures of image-based virtual try-on usually consist of clothes image generation and try-on image synthesis. In contrast, prior arts cannot guarantee satisfying clothes results when facing... 

In this project, a method is emoloyed to implement a Viterbi decoder on GPGPU. This method is based on combining all steps of the algorithm. This combination has some challenges that are related to differences between different steps of the algorithm. So in this project, some solutions are found to handle these challenges and a high-throughput Viterbi decoder is acheived  

The identification of the poor and the elimination of their poverty always has been the subject of discussion among researchers and policymakers. Today, most researchers consider poverty as a multidimensional phenomenon, and in addition to income indicator, it has indicators related to the level of knowledge, awareness, well-being and the health of individuals. The study of multidimensional poverty and the search for the discovery of its static and dynamic relationship with income-based poverty is the subject of this research. The analysis shows that both statically and dynamically, there is a substantial mismatch in the identification of the poor households according to these two... 

Cardiovascular diseases are the leading cause of death globally so early diagnosis of them is important. Many researchers focused on this field. First signs of cardiac diseases appear in the electrocardiogram signal. This signal represents the electrical activity of the heart so it’s primarily used for the detection and classification of cardiac arrhythmias. Permanent monitoring of this signal is not possible for specialists so we should do this by means of Artificial Intelligence. In this thesis, we use recurrent neural networks to classify electrocardiogram’s arrhythmias. This deep learning method, use two sources of data to learn from. The first part of data is global for everyone and the... 

It is well known that the classical continuum theory does not consider size effect in predicting material’s behavior, however at the micro and nano scales, this effect is not negligible and the classical theory has deficiencies at these scales. To take into account the size effect, higher order continuum theories introduce new material constants into the formulation. One famous version of these theories is couple stress which is applied in this research. This theory considers size effect of inhomogeneity by employing Characteristic length of materials. Quasi-static couple stress and classical dynamic approaches do not include explicitly the size of the material unit cell in the formulation... 

With the advance of the semiconductor technology, it has been feasible now to fabricate high performance multiprocessor systems-on-chips, containing tens and in the near future hundreds of processing cores. Meanwhile, the diminished technology feature size and high power consumption limit in using previous communication structures such as shared buses. Networks-on-Chip (NoC) have been introduced to obviate these constraints by providing high modularity, scalability and inherent parallelism. Design of these interconnect structures includes multiple tradeoffs between delays, throughput, energy consumption and silicon area requirements. So far, many mechanisms have been proposed to improve the... 

This study strives to solve the course scheduling problem based on the constraints in Sharif University. In course scheduling problem, first we need to assign each course to a professor (This is done according to professors’ preferences in Sharif University) then we assign each course to a classroom and a time slot subject to professors, classrooms and time constraints (For example no two courses can be assigned to the same classroom at the same time slot). If solved manually, the solving process will take several days and at best results in a feasible solution. The education department is responsible for scheduling 661courses in five days a week and assigning them to 56 classrooms... 

Low-frequency oscillation damping in power system is main determinants of the stability of the system. According to increasing of wind farms with high penetration levels in the power system, participation and control of wind farms for increase damping inter area oscillations seems important. In this study a damping controller suggested for Doubly Fed Induction Generator (DFIG). Using this controller and only available local signal as input controller and combination of both active and reactive power modulation, can greatly increase damping inter area oscillations and stability of uncertainty and different working points. Using linear analysis system, identified modes oscillating system and... 

Wind energy, as a type of renewable energy resource, is clean and is rapidly growing globally. The intermittency in the production of wind energy is the most major obstacle for these producers in a competitive electricity market, to compete with thermal units. Most existing models in the literature mitigate the risk of wind power, by coordinating them with energy storages such as pumped storages power plant. Here, in this thesis, another risk mitigation approach is introduced which combines wind energy and natural gas power plant in an electricity market. In addition, as the penetration level of wind power grows, the wind power producers must be considered as a price-maker player which their... 

Designing a part of an on site repair system for refurbishing steam turbine high pressure control valve has been done in this project. The main role of this control valve is to control the input flow to the turbine in different working condition . However, due to high temperature and pressure, the seat is exposed to erosion. R epairing this part in workshop requires great amount of time and cost, because the valve is a part of the turbine body and must be cut and rewelded to the turbine body. S o using an on-site repair method is highly needed. In addition, due to the limited visibility and access to the valve seat, and accurate... 

Fractured reservoirs have allocated a significant portion of Iran's and the world's oil and gas reserves, underscoring the importance of predicting the performance of these reservoir types in the oil and gas industry. Predicting the performance of fractured reservoirs by professional reservoir simulators, especially in enhanced recovery conditions, requires precise information on fracture flow functions, which is often limited. These functions vary for each reservoir based on the distinct characteristics of the rock and fluid, and the only way to determine them is through appropriate laboratory experimentation and modeling of the results. Currently, in simulation models, the relative... 

The study of evaporation of fuel droplet and determination of the rate of vaporization are important in designing combustion chambers. For achieving high performance of a combustor, the evaporation of fuel droplets takes place within a high pressure environment. At these conditions, the use of low-pressure models is not appropriate and many effects that are assumed negligible at low ambient pressures become very important. For example, the solubility of the ambient gas into the liquid phase is increased by increasing the ambient pressure. In addition, the ideal gas assumption is not valid for these conditions and one should use an appropriate equation of state (EOS) that can predict the...