Sharif Digital Repository

The goal of this thesis is using sparse decomposition to design secure steganography algorithms to insert secret data into the different cover signals. We propose data embedding in higher semantic levels of the cover signal to reach this goal. Sparse decomposition represents a signal as a linear combination of its structural elements. Data insertion into the sparse coefficients slightly changes the effect of structural elements in the signal representation. So, quality of the stego signal is preserved and such a steganography method leads to higher imperceptibility. In addition, data insertion into the higher semantic levels leads to the higher undetectability. Steganalyzers use statistical... 

Before performing any operation on a speech signal, it is necessary to properly remove the environmental noise existing on it. Noise Canceling Operation on Speech Signal is called speech enhancement. Up to now, many studies have been conducted on various ways to enhance the speech signal. Among the existing methods, statistical methods have proven to be superior to others. In all noise removal methods, the main challenge is that most noises are non-stationary. Since most of the noises in the environment are non-stationary, we are still looking for the better ways to remove them. With the advent of deep neural networks and their successful results in areas such as machine learning, a method... 

Many companies feel the need of automatic person identification for tracking entrance and exit of their employers. This system will help the company in reducing security costs, waste of time and increasing security. In this thesis a novel idea for 3d convolutional neural network text independent speaker verification system will introduced. Since many scenarios visionary data like lip motion, besides speech data, is available, in this thesis both of these data streams are used for developing designed speaker verification system. Multimodalily helps the verification system to be more robust to the noise and attacks. One of the main challenges in designing multimodal neural network speaker... 

The use of functional electrical stimulation after spinal cord injury is one of the most effective ways to prevent muscle atrophy.” Enzymes construction set, torque output increases and improved fatigue resistance”.
Functional electrical stimulation devices help in the operation of a muscle by applying a weak electrical current. This stimulation causes muscle contraction. In other words, it replaces the neural signals required to move the limb that cannot be produced due to causes such as spinal cord injury (SCI) orsecondary problems that arise after SCI.
One of the drawbacks of using FES is jerky motions in the muscles and limps. Such movements are not only harmful, but results in... 

Nowadays due to increasing environmental considerations and effort to reducing of the fossil fuels consumption, the necessity of hydroelectric power plants development is emerged. The purpose of this study is to suggest a method of spatial analysis for finding sites with appropriate potential for construction of these power plants, using TOPSIS and GIS. In this research, using the Python programming language, in software environment of Arc GIS version 9.3, local hydrological model was generated using the DEM layer and using of streams ordering methods, the overall situation along the river were processed. In the next step, with using of fuzzy TOPSIS algorithm implementation in software... 

Deep neural networks have not progresses comparative until last decade due to computational complexity and principal challenges as gradient vanishing. Thanks to newly designed hardware architecture and great breakthroughs in 2000s leading to the solution of principal challenges, we currently face a tsunami of deep architecture utilization in various machine learning applications. Sparsity of a representation as a feature to make it more descriptive has been considered in different deep learning architectures leading to different formulations where sparsity is impose on specific representations. Due to the gradient based optimization methods for training deep architecture, smooth regularizers... 

In low energy limit, some string theory models are described with an effective lagrangian which consist of two term. first is Eistein-Hilbert term and second is chern-simons. Adding the chern-simons term to Eistein-Hilbert lagrangian leads to the new modified field equations which schuwarzshild, Reissner-Nordstrom and FRW metrics satisfy these new modified equations while kerr metric is not a solution for these new fields. Since the gravitational field of spinning objects are similar to electromagnetic field, we expect to observe the chern-simons effects in gravitomagnetic component of the gravitational field. Here, in order to better understanding of subject, we investigate the Maxwell... 

Blind image steganalysis is a technique used to, which require no prior information about the steganographic method applied to the stego im- age, determine whether the image contains an embedded message or not. The basic idea of blind steganalysis is to extract some features sensitive to information hiding, and then exploit classifiers for judging whether a given test image contains a secret message.The main focus of this research is to design an choose features sen-sitive to the embedding changes. In fact, we use high order moments in different domains, such as spatial, DCT and multi-resolution do-main, in order to improve the performance of existing steganalyzers.Accordingly, First, we... 

Steganography is the art and science of writing hidden messages in such a way that no one, apart from the sender and intended recipient, suspects the existence of the message, a form of security through obscurity. In this thesis, we focus on entropic issue of multimedia signal in the two branches of Information hiding namely Steganography and Watermarking. How to choose the block and noise estimation in the watermarking, and analysis of the singular values decomposition in steganography are examples of using entopic issue which we use in our thesis. The two new designs for video signals AVI are presented in Watermarking. For the both proposed method ,first AVI video signal will be divided... 

Steganalysis is the science of detecting covert communication. It is called blind (universal) if designed to detect stego images steganographied by a wide range of embedding methods. In this method, statistical properties of the image are explored, regardless the embedding procedure employed. The main problem for image steganalysis is to find sensitive features and characteristics of the image which make a statistically significant difference between the clean and stego images. In this thesis we propose a blind image steganalysis method based on the singular value decomposition (SVD) of the discrete cosine transform (DCT) coefficients that are revisited in this work in order to enhance the... 

This thesis explored and described a new steganalysis system modeling. Image steganalysis systems are divided into two parts: feature extraction from images and classification of images. Many researches have been done in both parts and satisfactory results have also been reported. There are acceptable steganalysis methods which work accurately in high rates of steganography; however steganography in low rates is still undetectable. By lowering the rate of steganography in images, the difference between stego images and clean images would be reduced which accordingly led to the reduction of the difference between the corresponding extracted features. Thus, the ability of classification... 

In this project we proposed a new method for improving the detection abality of a steganalyser with a pre-processing on contents of an image. Steganalysis, using machine learning, is designing a classifier with two classes: Stego or Cover. This classifier should be trained with extracted features from signal. The result of the training procedure is a machine that decides a signal belongs to stego or cover class. The first step of steganalysis process is extraction of proper features from signal. Proper feature is a variable that represents all of the useful properties of signal. Second step of this process is classifying data to two class of stego and cover. Many algorithms are proposed for... 

The rapid increase in the size of digital multimedia data collections has resulted in wide availability of multimedia contents to the general users. Effective and efficient management of these collections is an important task that has become a focus in the research of multimedia signal processing and pattern recognition. In this thesis, we address the problem of automatic classification of music, as one of the main multimedia signals. In this context, music genres are crucial descriptors that are widely used to organize the large music collections. The two main components of automatic music genre classification systems are feature extraction and classification. While features are a compact... 

This project aims at forgery detection in digital videos. Most of existing methods are based on similar methods for image forgery detection. Therefore, they do not have sufficient accuracy in case of video forgery detection. In this project, we focus on local copy/move attacks in digital videos and propose 3 solutions for 3 problems in this field: 1) detection of copy/move along time axis (temporal copy/move), 2) detection of copy/move along x and y axes (spatial copy/move) and 3) detection of original and fake part in case of finding a duplication. For each of these 3 problems a feature extraction algorithm and a forgery detection algorithm are proposed. Feature extraction algorithms are...