A. Mohamed

Implementation of image fusion techniques for multi-focus images using FPGA

Image fusion is a process which combines the data from two or more source images from the same scene to generate one single image containing more precise details of the scene than any of the source images. Among many image fusion methods like averaging, principle component analysis and various types of Pyramid Transforms, Discrete cosine transform, Discrete Wavelet Transform special frequency and ANN and they are the most common approaches. In this paper multi-focus image is used as a case study. This paper addresses these issues in image fusion: Fused two images by different techniques which present in this research, Quality assessment of fused images with above methods, Comparison of different techniques to determine the best approach and Implement the best technique by using Field Programmable Gate Arrays (FPGA). First a brief review of these techniques is presented and then each fusion method is performed on various images. In addition experimental results are quantitatively evaluated by calculation of root mean square error, entropy; mutual information, standard deviation and peak signal to noise ratio measures for fused images and a comparison is accomplished between these methods. Then we chose the best techniques to implement them by FPGA.

Fast digital watermarking techniques for still images

In this paper, we investigate the problem of choosing the best domain for embedding watermarks in digital still images. During this paper, twelve watermarking techniques were implemented and evaluated. Four techniques were selected to represent different approaches of embedding data in spatial domain, three approaches using discrete cosine transform (DCT) domain, two approaches using discrete wavelet transform domain (DWT) domain, and two combined DWT-DCT techniques. The algorithms were chosen to represent a range of computational complexities and implementation structures. The performance of the selected algorithms was evaluated with respect to many perceptibility parameters.

Performance simulation of IEEE 802.16e WiMAX physical layer

Hardware implementation of WiMAX systems using field programmable gate array (FPGA) technology is a very important task for mobile communication and wireless network researchers. As a first step, a careful simulation of the system performance has been done using the simulation link platform of matlab_2009. The performance of the simulated system was evaluated using the bit-error-rate (BER). In this part; we have discussed different algorithms of the orthogonal frequency division multiplexing (OFDM), as well as the WiMAX physical layer based on the IEEE 802.16e standard under different combinations of digital modulation schemes and different communication channels. The simulation results show that adaptive modulation achieves similar spectral efficiency with better energy efficiency. When channel conditions are poor, energy efficient schemes such as BPSK or QPSK were used. As channel quality improves, higher M-ary modulation schemes such as16-QAM or 64-QAM was used.

An efficient algorithm in extracting human iris Morphological features

The interface of computer technologies and biology is having a huge impact on society. Human recognition research projects promises new life to many security-consulting firms and personal identification system manufacturers. Iris recognition is considered to be the most reliable biometric authentication system. Very few iris recognition algorithms were commercialized. The method proposed in this paper differed from the existing work in the iris segmentation and feature extraction phase. Digitized grayscale images of Chinese Academy of Sciences—Institute of Automation (CASIA) database were used for determining the performance of the proposed system. The circular iris and pupil of the eye image were segmented using Morphological operators and Hough transform. The localized iris region was then normalized in to a rectangular block to account for imaging inconsistencies. This method provides accurate features as well as simple and fast iris analysis methods.

Automated algorithm for iris detection and code generation

The interface of computer technologies and biology is having a huge impact on society. Human recognition research projects promises new life to many security-consulting firms and personal identification system manufacturers. Iris recognition is considered to be the most reliable biometric authentication system. Very few iris recognition algorithms were commercialized. The method proposed in this paper differed from the existing work in the iris segmentation and feature extraction phase. Digitized grayscale images of Chinese Academy of Sciences-Institute of Automation (CASIA) database were used for determining the performance of the proposed system. The circular iris and pupil of the eye image were segmented using morphological operators as well as circular Hough transform. The localized iris region was then normalized into a rectangular block to account for imaging inconsistencies. Finally, the corresponding iris code was generated using the Euclidean distance. This method provides accurate signature as well as accurate code of the human iris in a simple and fast way.

Automated classification techniques of galaxies using artificial neural networks based classifiers

Processing and classifying galaxy information is one of the most important challenges and intensive research area for astronomers. In this paper; analyzing and classifying photographic images of galaxies are presented, with interesting scientific findings gleaned from the processed photographic data. In addition, the performance of ten artificial neural networks (ANNs) based classifiers was evaluated, based on a selected set of features. They are a combination of a set of morphic features; derived from image analysis and principal component analysis (PCA) features. These features are combined and arranged to constitute five groups of features. The results showed that; the support vector machine (SVM) based classifier provides the best results; about 99.529% for a feature set composed of the nine morphic features and 24 principal components; occupying 85% of the original data. The dataset was ten cases of NGC category taken from standardized catalog from Zsolt Frei website.

Energy efficient collaborative proactive routing protocol for Wireless Sensor Network

Abstract A Wireless Sensor Network (WSN) is a group of tiny power-constrained nodes that cover a vast region of interest (ROI), sense and communicate it to the Base Station (BS). The main challenge encountered in WSNs is how to cover the ROI perfectly and transmit the monitored data to the BS for the longest possible time. Although many energy-efficient routing protocols for periodic monitoring applications were recently introduced, the dynamic nature and complex environments of WSN applications make building such protocols a considerable challenge. In this paper, the node degree of the Degree Constrained Tree (DCT) in homogeneous proactive WSN is studied analytically for the network with one BS that is outside the ROI. Since the node degree affects the network lifetime of these types of networks, the optimum node degree for minimum energy consumption in DCT is derived. Subsequently, the paper proposes a Collaborative Distributed Antenna (CDA) routing protocol that is based on distributed antenna theory to provide fair load distribution in terms of transmission energy. CDA is based on DCT with optimal node degree and is designed for periodic data monitoring in WSN applications. The experimental results prove our analysis to emphasize that using optimal node degree in DCT doubles its network lifetime compared to using other node degrees. Moreover, adding CDA to DCT with optimal node degree is proved to double the network stability period and reduce the ratio between instability period and the network lifetime to its half. It also shows 25% increase in network lifetime and minimum rate of node loss compared to its peers, such that the lifetime of half the nodes is preserved until few rounds before the end of network lifetime.

Implementation of fast digital watermarking techniques

In this paper, we investigate the problem of choosing the best domain for embedding watermarks in digital still images. During this paper, eight watermarking techniques were implemented and evaluated. Four techniques were selected to represent different approaches of embedding data in spatial domain, three approaches using discrete cosine transform (DCT) domain, and one approach using discrete wavelet transform domain (DWT) domain. The algorithms were chosen to represent a range of computational complexities and implementation structures. The performance of the selected algorithms was evaluated with respect to many perceptibility parameters.

Joint 2D-DOA and Carrier Frequency Estimation Technique Using Nonlinear Kalman Filters for Cognitive Radio

The problem of jointly estimating carrier frequencies and their corresponding two-dimension direction of arrivals (DOA) of band-limited source signals is considered in this paper for cognitive radio. The main problem of estimating carrier frequencies spread over a wideband spectrum is the requirement of high sampling rates. Thus, the Kalman filters are applied in the spatial domain instead of the temporal domain in the proposed algorithm to relax hardware complexity. The proposed algorithm exploits both the azimuth and elevation angles instead of a single DOA to increase the spatial capacity. Two approaches are proposed using two different types of nonlinear Kalman filter: extended Kalman filter (EKF) and unscented Kalman filter (UKF). Using simulations, the factors that affect the performance of both the filters are discussed. Scaling the estimated parameters to the same range and the proper tuning and initialization of the filters are crucial factors to prevent the filter divergence. Although UKF is supposed to have a better performance than EKF, reducing the inter-element spacing of the employed arrays and the proper filter initialization can make EKF approach the performance of UKF. On the other hand, UKF suffers from high processing time. Overall, both filters are able to converge to the true values of the unknown parameters using a number of relaxed analog-to-digital converters equal to the number of the array elements in the employed arrays. However, the approaches can detect a number of source signals higher than one-third of the total number of the array elements.

Development of Iris Security System Using Adaptive Quality-Based Template Fusion

Recently, the interface of computer technologies and biology has an enormous impact on society. Human recognition research projects promise a new life to various security consulting. Iris recognition is considered to be one of the exceedingly reliable authentication systems. To account for iris data variations, nearly of all iris systems store multiple templates per each user. Approaching to overcome the storage space and computation overheads, this paper proposes an intelligent fusion technique, algorithms, and suggestions. The quality of the input image has been checked firstly to ensure that “qualified iris samples” will be only treated. The proposed system, with the aid of the image selection stage and template quality test, has the advantage of being adaptive and simple, but can come at the expense of reject extremely inadequate data for any user. Complete the eye shape using the convex hull is the main key for image selection module. Shape-based thresholding is integrated with morphological features to avoid the dark iris problems, elliptical pupil and iris shapes. For best recognition rate, the optimum values of the 1D log Gabor filter parameters and different sizes of the iris code are recorded. From experimental results, an HD value of 0.4 can be chosen as a suitable separation point, and the optimum code size was found to be 20 × 480. An experimental work reveals a reduction in database size by nearly a 78 % and an increase of verification speed of about 85.25 % is achieved while 14.75 % of computation time in the shifting process is only required. Comparing with existing algorithms, the proposed algorithm gives an accuracy of 96.52 and 99.72 % for TRR and TAR, respectively for closed loop tested dataset. In addition, with a number of experimentations, the proposed algorithm gives the lowest FAR and FRR of 0.0019 and 0.00287 % respectively and EER of 0.0024 %.