Sharif Digital Repository

According to needs for high quality pictures with high compression, in variety of applications such as Satellite communications, in this thesis some parts of JPEG2000 standard with computation and time complexity, have been implemented on FPGA. This parts are:
1- Two dimensional Discrete wavelet transform (2D-DWT)
2- Bit Plan Coding
3- Context-based Adaptive Binary Arithmetic Coding (MQ-Coder)
Each of these parts is implemented on basis of the best architectures in papers with some improvement on its hardware; and the results are compared with other similar works. Other parts, such as “Packet Creation” that are intrinsically software-based, are not implemented and it is... 

In this thesis a dynamical model is introduced for Phonocardiogram based on its morphology which is capable of generating synthetic signals with realistic morphology. The model is for normal Phonocardiograms which includes the two dominant heart sounds, namely the first and the second sounds, and is inspired by the Electrocardiogram dynamical model. In the proposed dynamical model each heart sound is modeled with several Gabor kernels. The ultimate goal of such dynamical model is to establish a model-based processing framework for Phonocardiogram. This framework is devised by employing the dynamical model equations within an Extended Kalman Filter structure, and the simultaneously recorded... 

Finemet soft magnetic alloys in the form of toroidally winded ribbons are not suitable for industrial applications where a large volume of magnetic materials is required. Production of Finemet bulk alloy by powder metallurgy techniques is an applicable method to produce complex component with isotropic magnetic properties which are the same as ribbons. In this research, Finemet bulk magnetic alloy with composition of has been produced by consolidation of amorphous powders obtained by milling of melt-spun ribbons. At the all stages, the structure and magnetic properties were studied using X-ray diffraction, differential scanning calorimetry, transmission electron microscopy, scanning... 

In present research forming process of nanopowders, which is a part of powder metallurgy was investigated by molecular dynamics method. Powder metallurgy is a relatively new method for production of industrial parts by pouring powder into die and compaction to desired density. One can reach parts with higher quality and strength by decreasing size of powder’s particles and entering the nano scale. Particle with smaller size have higher specific surface and due more intensity to react. Classic methods for investigation of this process don’t cover the atomic scale effects, so using newer procedures such as molecular dynamics is highly recommended. In present research, at first compaction of... 

Wireless Multimedia Sensor Networks (WMSNs) have many applications in survilient systems. In this kind of applications, although we have limited energy and computational power, we should provide minimal security to transfer data in such networks. These constraints are more crucial when we use multimedia that has more data for processing and transmission. Because of this huge amount of data, we must use compression for using less energy in both processing and transmission.
In some algorithms, it has been proposed to integrate compression and encryption in order to decrease time and power consumption. The proposed algorithm is several times faster than AES (Advanced Encryption Standard) or... 

One of the funadmental problems in signal processing is finding the best sampling strategy of a continuous time signal, and then finding the best strategy for compressing the obtained signal samples. The sampling and compression steps are designed with the aim of minimizing the distortion of the reconstructed signal. The problem of finding the best sampling strategy has been widely studied in the signal processing literature. In particular, the Nyquist–Shannon sampling theorem gives a sufficient condition for perfect reconstruction of a continuous-time signal of finite bandwidth. Most of the signal processing literature deals with deterministic signals, with relatively less attention paid to... 

In present research forming process of nanopowders, which is a part of powder metallurgy was investigated by molecular dynamics method. Powder metallurgy is a relatively new method for production of industrial parts by pouring powder into die and compaction to desired density. One can reach parts with higher quality and strength by decreasing size of powder’s particles and entering the nano scale. Particle with smaller size have higher specific surface and due more intensity to react. Classic methods for investigation of this process don’t cover the atomic scale effects, so using newer procedures such as molecular dynamics is highly recommended. In present research, at first compaction of... 

Payload Attribution Systems (PAS) are one of the most important tools of network forensics for detecting offenders and victims after the occurrence of a cybercrime. A PAS stores the network traffic history in order to detect the source and destination pair of a certain data stream in case a malicious activity occurs on the network. The huge volume of information that is daily transferred in the network means that the data stored by a PAS must be as compact and concise as possible. Moreover, the investigation of this large volume of data for a malicious data stream must be handled within a reasonable time. For this purpose, several techniques based on storing a digest of traffic using Bloom... 

The present research focuses on the simulation of the metallic powder compaction process with the Peridynamics theory. Various methods are exploited to simulate this process in literature. Studying the nonlinear behavior of powders includes different phenomena such as dislocation and grain boundary, making it complicated. However, numerous research has been shaped to consider these phenomena on the micro-scale. There is also another batch of nano-scale studies underway. In this class of simulations, considering atoms as rigid particles, interatomic potentials, and molecular dynamics methods are used. Because of atomic-scale precision, this approach has very high accuracy. The massive... 

Nowadays, one of the challenges in the field of bioinformatics is the excess processed data volume such that this data volume resulted from a complete genome sequence of a species can be up to hundreds gigabytes. Every time that we talk about increasing data volume, data storage, transforming, and the process will become of interest. Moreover, considering the presence of portable sequencer devices in the market and the limitations of process outside of the lab environments, this problem becomes of more critical importance. Fortunately, due to the nature of the genome data and their redundancy, specific algorithms to compress them have been introduced to the market. In this thesis, we chose... 

Recently, portable genome sequencing devices have been introduced to the market, which have also made it possible to provide these services in remote locations or outside the laboratory. The amount of raw data from the readings of a sequencer for the entire genome of a human or plant can be in the hundreds of gigabytes, making it difficult and expensive to maintain and transfer to the center for such sequencing. Fortunately, these readings have a lot of redundancy, and many new algorithms have been proposed to compress them based on the intrinsic properties of this data. Sequencing devices were mainly used in the laboratory environment, which naturally had virtually unlimited access to urban... 

Nowadays sharing data between users and organizations can be difficult due to privacy and legal reasons. Therefore, real-world data is not fully exploited by machine learning methods. A novel model training method called Federated Learning was proposed to alleviate the aforementioned problem by enabling users to jointly train learning models without sharing data. Federated learning utilization faces many challenges among which high communication cost, statistical heterogeneity and system heterogeneity are most important. This research deals with these challenges by proposing two methods of Reduced Clustering Federated Learning (RCFL) and Weighted Federated Distillation (WFD). Moreover, we... 

Nowadays, neural networks are used as the main method in most machine learning applications. But research has shown that these models are vulnerable to adversarial attacks imperceptible changes to the input of neural networks that cause the net- work to be deceived and predict incorrectly. The importance of this issue in sensitive and security applications of neural networks, such as self-driving cars and medical diagnosis systems, becomes much higher. In recent years, many researches have been done in the field of making neural net- works robust against this threat, but in most of them, higher robustness has been provided on the basis of larger and more complex models. Few researches have... 

In recent years, the use of Deep Neural Networks in solving various problems has grown considerably. Possessing a large number of parameters, these networks have the ability to reconstruct complex functions and relations from large amounts of data and have been able to achieve the best results in a wide range of problems. But using these models comes with its own problems. These networks typically require considerable resources in order to run. This makes it inefficient or impossible to use them in systems with limited processing capabilities, e.g mobile phones. The existing approaches, e.g. the deployment of the model on a powerful server and network compression, have their own drawbacks...