Sharif Digital Repository

In this paper, we propose an information theoretic approach to fuse images compressed by compressed sensing (CS) techniques. The goal is to fuse multiple compressed images directly using measurements and reconstruct the final image only once. Since the reconstruction is the most expensive step, it would be a more economic method than separate reconstruction of each image. The proposed scheme is based on calculating the result using weighted average on the measurements of the inputs, where weights are calculated by information theoretic functions. The simulation results show that the final images produced by our method have higher quality than those produced by traditional methods, especially... 

Deep neural networks have been shown to be vulnerable to adversarial examples deliberately constructed to misclassify victim models. As most adversarial examples have restricted their perturbations to the Lp-norm, existing defense methods have focused on these types of perturbations and less attention has been paid to unrestricted adversarial examples; which can create more realistic attacks, able to deceive models without affecting human predictions. To address this problem, the proposed adversarial attack method generates an unrestricted adversarial example with a limited number of parameters. The attack selects three points on the input image and based on their locations transforms the... 

Segmentation plays an important role in the machine vision field. Extraction of dominant segments with large number of pixels is essential for some applications such as object detection. In this paper, a new approach is proposed for color image segmentation which uses ideas behind the social science and complex networks to find dominant segments. At first, we extract the color and texture information for each pixel of input image. A network that consists of some nodes and edges is constructed based on the extracted information. The idea of community detection in social networks is used to partition a color image into disjoint segments. Community detection means partitioning vertices of a... 

A new method for image registration has been previously proposed by the authors, which the registration is based on physical forces. The registration parameters are translation and rotation. This method assumes images like charged materials that attract each other. In this case, one of the images moves in the same direction as the applied force while the other one is still. The movement of the image continues until the resultant force becomes zero. This approach estimates the registration parameters simultaneously and leading to a better optimized set of registration parameters. The registration error for this method is 1 to 3 pixels. In this paper we aim to develop this method for the... 

This paper presents a new segmentation method for color images. It relies on soft and hard segmentation processes. In the soft segmentation process, a cellular learning automata analyzes the input image and closes together the pixels that are enclosed in each region to generate a soft segmented image. Adjacency and texture information are encountered in the soft segmentation stage. Soft segmented image is then fed to the hard segmentation process to generate the final segmentation result. As the proposed method is based on CLA it can adapt to its environment after some iterations. This adaptive behavior leads to a semi content-based segmentation process that performs well even in presence of... 

Vivid banding patterns in medical images of chromosomes are a vital feature for karyotyping and chromosome classification. The chromosome image quality may be degraded by many phenomenon such as staining, sample defectness and imaging conditions. Thus, an image enhancement processing algorithm is needed before classification of chromosomes. In this paper, we propose an adaptive and iterative histogram matching (AIHM) algorithm for chromosome contrast enhancement especially in banding patterns. The reference histogram, with which the initial image needs to be matched, is created based on some processes on the initial image histogram. Usage of raw information in the histogram of initial image... 

The goal of image segmentation is to find regions that represent objects or meaningful parts of objects. In this paper a new method is presented for color image segmentation which involves the ideas used for community detection in social networks. In the proposed method an initial segmentation is applied to partition input image into small homogeneous regions. Then a weighted network is constructed from the regions, and a community detection algorithm is applied to it. The detected communities represent segments of the image. A remarkable feature of the method is the ability to segments the image automatically by optimizing the modularity value in the constructed network. The performance of... 

In this paper, we propose two lossless compression techniques that represent a two dimensional Run-length Coding which can achieve high compression ratio. This method works by partitioning the block regions of the input image into rectangles instead of working by runs of adjacent pixels, so it is found to be more efficient than 1D RLE Run-length Coding for transmitting texts and image. In the first method, length and width of consecutive black and white rectangles are transmitted. The idea of this method is new and it can be very effective for some images which have large blocks of black or white pixels. But in the second method only black rectangles are considered in order to transmit and... 

In this paper we present an adaptive method for image compression that is based on complexity level of the image. The basic compressor/de-compressor structure of this method is a multilayer perceptron artificial neural network. In adaptive approach different Back-Propagation artificial neural networks are used as compressor and de-compressor and this is done by dividing the image into blocks, computing the complexity of each block and then selecting one network for each block according to its complexity value. Three complexity measure methods, called Entropy, Activity and Pattern-based are used to determine the level of complexity in image blocks and their ability in complexity estimation... 

This paper presents a complexity-based image compression method using neural networks. In this method, different multi-layer perceptron ANNs are used as compressor and de-compressor. Each image is divided into blocks, complexity of each block is computed using complexity measure methods and one network is selected for each block according to its complexity value. Three complexity measure methods, called Entropy, Activity and Pattern-based are used to determine the level of complexity in image blocks and their ability are evaluated and compared together. Selection of a network for each image block is based on its complexity value or the Best-SNR criterion. Best-SNR chooses one of the trained... 

The effects of real and Gleeble simulated double pass thermal cycles on the properties of the intercritically reheated coarse grained heat affected zones in X80 microalloyed pipeline steel has been investigated. The Gleeble simulated process involved heating the X80 steel specimens to the first peak temperature of 1400°C and then reheating to the second peak temperature of 800°C, with different cooling rates. The size and area fraction of martensite/austenite (M/A) constituents were obtained by a combination of field emission scanning electron microscopes and image analysis software. In addition, misorientation was characterized by electron back-scatter diffraction analysis. It is clear that...