Hossam M. A. Fahmy

Expression and illumination invariant preprocessing technique for Face Recognition

Face recognition is one of the most intensively studied topics in the field of computer vision and pattern recognition. Facial expression, which changes face geometry, usually has an adverse effect on the performance of a face recognition system. Also, the appearance of a face image is usually affected by illumination conditions that hinder the automatic face recognition process. In this paper, we propose a robust expression and illumination invariant preprocessing technique for face recognition. We propose a new wavelet based method to eliminate the effect of expression variation and combine this method with an existing DCT based method for illumination normalization. Experimental results on the Yale database show that the proposed approach improves the face recognition rates.

Illumination invariant face recognition in logarithm Discrete Cosine Transform domain

Face is considered to be one of the biometrics in automatic person identification. For face recognition system to be practical, it should be robust to variations in illumination, pose and expression as humans recognize faces irrespective of all these variations. In this paper, we present an illumination invariant face recognition method in the logarithm Discrete Cosine Transform domain. We use an existing illumination normalization technique in the logarithm DCT domain. The main contribution in this paper is that we apply the Principal Component Analysis (PCA) algorithm for feature extraction in the DCT domain. By this, we skip the inverse DCT step and reduce the computational cost. Experimental results on the Yale B database show that we obtain the same results exactly as applying PCA in the spatial domain with the advantage of the reduced computational cost.

RCA: Efficient connected dominated clustering algorithm for mobile ad hoc networks

Abstract Clustering of mobile ad hoc networks is a largely growing field. The perceived benefits of clustering are comprehensively analyzed in open literature. This paper considers the development of a new distributed connected-dominated-set clustering algorithm called Ring Clustering Algorithm (RCA). RCA is a heuristic algorithm that has three phases: ring-formation phase, members-joining phase and CDS-nodes selection phase. In the ring-formation phase, each ring consists of three ring-nodes. The ring is formed if it has the highest priority. The priority of the ring is based on the total ring-degree rather than the individual node-degree. The degree of a ring is the number of neighbors that the three ring-nodes have all together without repetition. Nodes that cannot form rings join neighboring rings as members in the members-joining phase. In the CDS nodes selection phase, the decision is made for a node to remain or leave the CDS. This paper presents the proof that the maximum number of rings that can be formed by RCA in any disk area equals the maximum number of independent nodes that create non-overlapping circles in a corresponding area. This allowed RCA to achieve the lowest fixed approximation ratio (5.146). Moreover, RCA has O(n) for both time and message complexities. Thus, RCA algorithm outperforms the current-best CDS algorithms that are investigated in this paper.

Modular Multiplication for Public Key Cryptography on FPGAs

All public key cryptosystems, though being highly secure, have a common drawback: They require heavy computational effort. This is due to the reliance on modular multiplication of large operands (1024 bits or higher). The same problem arises in data encryption/decryption and digital signature schemes. Examples of such cryptosystems are RSA, DSA, and ECC. Now considering embedded platforms for applications of smart cards and smart tokens, the overall time performance of the cipher system becomes very slow. This refers to the limited computational power of the embedded processors. This paper introduces an enhanced architecture for computing the modular multiplication of large operands. The proposed design can act as a co-processor for embedded general purpose CPUs. The proposed design is compared with three previous architectures depending on carry save adders and look up tables, and scoring 69 MHz of maximum frequency. Look up tables should be loaded with a set of pre-computed values. Our proposed architecture replaces both look up tables and pre-computations with an enhanced version of sign detection techniques. Considering 1024 bits architectures, the proposed design scored a maximum frequency of 181 MHz. It also has a better overall absolute time for a single operation.

Agent-based trusted on-demand routing protocol for mobile ad-hoc networks

Abstract The routing performance in mobile ad hoc networks (MANETs) relies on the co-operation of the individual nodes that constitute the network. The existence of misbehaving nodes may paralyze the routing operation in MANETs. To overcome this behavior, the trustworthiness of the network nodes should be considered in the route selection process combined with the hop count. The trustworthiness is achieved by measuring the trust value for each node in the network. In this paper, a new protocol based on self monitoring (agent-based) and following the dynamic source routing (DSR) algorithm is presented. This protocol is called agent-based trusted dynamic source routing protocol for MANETs. The objective of this protocol is to manage trust information locally with minimal overhead in terms of extra messages and time delay. This objective is achieved through installing in each participated node in the network a multi-agent system (MAS). MAS consists of two types of agents: monitoring agent and routing agent. A new mathematical and more realistic objective model for measuring the trust value is introduced. This model is weighted by both number and size of routed packets to reflect the “selective forwarding” behavior of a node. The performance evaluation via simulation shows that our protocol is better than standard and trusted DSR. The simulation is done over a variety of environmental conditions such as number of malicious nodes, host density and movement rates.

PhishBlock: A hybrid anti-phishing tool

Phishing is a means of obtaining confidential information through fraudulent websites that appear to be legitimate. Anti-phishing detection techniques are either lookup based or classifier based. Lookup based systems suffer from high false negatives while classifier systems suffer from high false positives. To better detect fraudulent websites, we propose in this work an efficient hybrid system that is based on both lookup and a support vector machine classifier that checks features derived from websites URL, text and linkage.

Efficient implementation of modular multiplication on FPGAs based on sign detection

All public key cryptosystems, though being highly secure, have a common drawback: They require heavy computational effort. This is due to the reliance on modular multiplication of large operands (1024 bits or higher). The same problem arises in data encryption/decryption and digital signature schemes. Examples of such cryptosystems are RSA, DSA, and ECC. Now considering embedded platforms for applications of smart cards and smart tokens, the overall time performance of the cipher system becomes very slow. This refers to the limited computational power of the embedded processors. This paper introduces an enhanced architecture for computing the modular multiplication of two large numbers X and Y modulo a given modulus M. The proposed design can act as a co-processor for embedded general purpose CPUs. The proposed design is compared with three previous architectures depending on carry save adders and look up tables. Look up tables should be loaded with a set of pre-computed values. Our proposed architecture uses the same carry save addition, but replaces both look up tables and pre-computations with an enhanced version of sign detection techniques. The proposed architecture supports higher frequencies than other architectures. It also has a better overall absolute time for a single operation.

ABRM: In-Network Aggregation Based Routing Protocol for Mobile Sensor Networks with Multiple Mobile Sinks

The recent technological advances in the field of wireless sensor networks (WSN) have expanded the range of WSN applications. In some of these applications, sensor nodes are mobile rather than static. Also, the recent advances in personal digitals assistants (PDAs) allow the existence of multiple mobile sinks to collect the sensors data. These characteristics require the design of new routing protocols to meet the existence of mobile sensors and multiple mobile sinks while taken into consideration the limited resources for sensor nodes especially energy. In-network aggregation is one of the important techniques used to save power consumption. This paper presents ABRM which is an in-network aggregation based routing protocol for mobile sensor networks with multiple mobile sinks. Compared to CCBR, ABRM yields good aggregation results in addition to a great reduction in power consumption and routing cost.

High Performance Java Card Operating System

Due to the fast evolving of trusted computing environments and internet-of-things an eager need has been established for open platforms which support interchangeable technologies to co-exist without threatening systems security. Certainly, future embedded applications will need high performance operating systems to support the intensive-computing algorithms required for satisfying acceptable response and secure the application inside the vulnerable open environment, hence, new inevitable requirements for embedded operating systems have arisen including hard real-time response, support for native applications, system openness and system scalability. This paper introduces a new design for secure and open smart card operating system, called ESCOS (Egypt Smart Card Operating System), based on the prevalent Java Card technology. The new design provides competitive characteristics in the main three factors of judging smart card platforms, namely, system security, supported technology and system response. In addition, ESCOS is designed to have high degree of modularity and re-configurability to meet fast-changing business needs and diverse hardware platforms.

Multilevel minimised delay clustering protocol for wireless sensor networks

In this paper, a multilevel minimised delay clustering protocol (MMDCP) is proposed. MMDCP is proved to extend the lifetime of wireless sensor networks through levelling and through a better choice of cluster heads. MMDCP assigns the number of the lower level cluster heads and the leaf nodes in the network so as to minimise the end-to-end delay. When comparing MMDCP to LEACH-C, THCHP and Delay-Aware protocols; MMDCP succeeds in extending the lifetime of the network and in minimising the end-to-end delay and also in increasing throughput values. MMDCP also improves the structure of the network of delay-aware protocol. Two radio models are used in evaluation, the first order radio model and the discrete radio model. The obtained results are proved analytically and via simulation. These results make the proposed protocol a very good candidate for use in crises management applications like pre-expectation of landslides and early control of slum fires.