S. Emalda Roslin

Text independent speaker recognition system using Back Propagation Network with wavelet features

Automated speaker recognition system is extremely important in areas such as Forensic and Defense. Performance of an automated speaker recognition system is dependent on feature extraction and classification. As speech is a non-stationary signal and the information is present in low frequencies, it necessitates a non-stationary tool that performs multi resolution analysis. An exhaustive approach is carried out in this work to identify the wavelet that is best suited for feature extraction. Of the various wavelets, Discrete Meyer Wavelet provides higher inter-class variance and lesser intra- class variance. Sixteen features are extracted for wavelet co-efficients and a five layered Back Propagation Network is used for recognizing the speakers.

Light weight cryptographic algorithms for medical internet of things (IoT) - a review

Internet of things (IoT) is communication between smart objects and human. It finds enormous applications in the field of healthcare monitoring, information management system, agriculture, predicting the natural disaster etc. In all those applications of IoT, security plays a vital role. In this paper, a study on various encryption light weight techniques used for IoT was analyzed. Also the performance of those existing algorithms is analyzed in detail based on the constraint and merits of the IoT. This detailed survey is made towards obtaining a tradeoff between security, cost and performance of IoT based application.

ART Network Based Text Independent Speaker Recognition System for Dynamically Growing Speech Database

Automated recognizing a speaker from the speech signals is the foremost application in forensics. Speaker recognition system involves two phases namely feature extraction and a classifier system. Features extracted from the speech signals are fed to an already trained classifier system that identifies the speaker. Major challenge occurs when the database is periodically updated which necessitates retaining the classifier with new set of exemplars includes the old and new datasets. As training the neural network is computationally intensive, Back Propagation system is not ideal for speaker recognition system (updation). Hence it necessitates an efficient speaker recognition system that doesn’t forget the old database but adjusts to the new set of data. In this paper an Adaptive Resonance Theory (ART) based speaker recognition system is proposed that is capable of functioning well even in the case of periodic updation.

Improving performance of delay constraint scheduling protocols for wireless sensor network

In Remote Sensor Arrange deferral and vitality are vital limitations. It is watched that postponement, parcel conveyance proportion, vitality, rate transmission issues are created when a bundle is sent from sources to goal. This paper manages issues of postponement and vitality. Also, it deals with a Course choice paradigm which depends on least hub postponement and vitality utilization. At the point when clog happens in a way, then the information transmission rate at Macintosh layer is lessened utilizing 802.11. The recreation comes about demonstrate that proposed convention is more productive than existing convention under the end to end delay.

Genetic algorithm based cluster head optimization using topology control for hazardous environment using WSN

Wireless Sensor Network (WSN) is widely used in recent years for the applications where human intervention is impossible. In case of nuclear power plant if any small delay occurs for data forwarding due to any node failure may results in severe disaster. Hence effective Topology Control is required to obtain an energy efficient sensor network even if any node fails. An efficient topology control using genetic algorithm based cluster head selection is presented in this paper. In this work, three tier sensor network architecture is developed and is comprised of Super Head nodes, Cluster Head nodes and Cluster Slave nodes. The results obtained for the developed three tier architecture is compared with two tier and one tier architectures. Residual Energy, Bandwidth and Memory Capacity are used as selection criteria. Quantitative analysis is also carried out to study the impact of N-tiers on the performance of the proposed algorithms. From the quantitative analysis of the proposed methodologies on the N-tier architecture with various node densities, it has been proved that two tier architecture provides reduced energy consumption compared to others.

Lifetime and Transport Delay Optimization in Presence of Delay in WSN

The wireless sensor network is fast-emerging field nowadays. The small sensor nodes in WSN compacts with functionalities sensing, processing, and communication. This technology has many applications such as home automation, health monitoring, traffic, security monitoring. The primary function of sensor nodes is to detect and capture the information and send to sink node. This transmission of data consumes minimum amount of energy to maximize lifetime of WSN. For reducing delays, we discuss the different approaches to maximize the life of WSN. This paper implements the BCMN/A protocol to achieves energy and delay efficiency both in intracluster and intercluster. In the proposed system, we make system more secure by sending data in the encrypted format using encryption algorithm and detect the delayed attack. This method improves the energy and delay efficiency of the network and enhances the network lifetime.

Detection and prevention of black hole and selective forwarding attack in clustered WSN with Active Trust

In the recent era, security is the major problem in sensor networks. Wireless sensor networks (WSNs) are mostly used for various real-world applications. However, WSNs face a lot of insider and outsider attacks, and it is complex to identify and protect towards insider attacks. Generally, an insider attack, in which the intruders choose several received data packets to drop, threatens the clustered WSNs. This situation has occurred because of the unattended clustered environments in the network. To overcome this problem, this paper proposes a trustable and secure routing scheme using two-stage security mechanism, and dual assurance scheme, for selecting the node and securing the data packet for WSNs. Both schemes are based on Active Trust to protect several kinds of attacks, such as black hole attack, and selective forwarding attack, during routing. Therefore, this paper identifies the trusted path and provides the secure routing paths using trust and Cuckoo search algorithm. Energy is the performance parameter utilized in the proposed scheme. The experimental result proves that proposed system provides the assurance to prolong the network lifespan and the probability of secure routing path in the network.

Transposition based symmetric encryption and decryption technique for secured image transmission through internet

The role of internet became a non replaceable one in our day to day life. Internet users also face various security threats such as eavesdropping and unauthorized access through intrusion. Internet users need to be protected and provided privacy. Network security and image encryption has become important and high profile issues. Various cryptographic techniques are already available in literature for secured data transmission. Image encryption is still a major area of research with extensive effort being spent in determining a standardized image encryption technique that is difficult for hackers to crack the image. Hence the proposed work aims at developing an effective and standardized image encryption technique. At the transmission side it involves the following steps: subdividing an image in to n number of sub images of p × q, finding a suitable permutation/combination to interchange the pixels, merging the sub images in to a single image and transmitting it. At the receiver side the received encrypted image is subdivided in to n number of sub images of size p × q, the original pixel positions are obtained and the original image is formed. The effectiveness of the technique is measured in terms of correlation index and entropy for different rules of permutation. Finally the most effective permutation is identified which results in lesser correlation index and stronger entropy. The computational complexity of the algorithm is also analyzed.

Super Strongly Perfect Graphs in the Selection of Cluster Heads in Wireless Sensor Networks

Wireless sensor network comprises of dense sensor nodes which are randomly deployed. Major challenges in WSN are limited battery source and computation capacity. Considerable research has been carried out in the area of maximizing battery lifetime by reducing the energy consumption. Once such proposed technique involves hierarchical topology control. Conventionally proposed algorithm for hierarchical topology control involves computationally intensive soft computing tools. It leads to higher energy consumption in the sink node. Hence it necessitates computationally less intensive technique for cluster head selection. In this paper, an efficient cluster head selection is proposed using minimal dominating set in Super Strongly Perfect (SSP) graph.

A reversible encryption technique for storing weld piece related information during digital archiving of industrial radiographs

Archival of digital radiographs is the growing need in weld industries. However in addition to radiographs, the information about the weld piece must also be stored for future documentation. However the information about the radiographs may be confidential and should not be stored as such. Also matching the radiographs with the corresponding radiographs is also a major area of concern. Hence it necessitates an effective encryption technique that hides the information about the weld piece in the radiograph itself. The proposed technique involves hiding the nature of the weld defect in the radiographs. The paper also proposes an extraction technique to effectively extract the signature image.