Search

With the recent explosion in the number of wireless communication technologies, the frequency spectrum has become a scarce resource. The need of the hour is an efficient method to utilize the existing spectrum and Cognitive Radio is one such technology that can mitigate the spectrum scarcity. In a cognitive radio system, the unlicensed secondary user accesses the spectrum allotted to licensed primary users when it lies vacant. To implement dynamic or opportunistic access of spectrum, secondary users perform spectrum sensing, which is a quintessential part of a Cognitive radio. From the Cognitive user’s point of view, lesser error probability means an increased likelihood of channel reuse when it is vacant, and a higher detection probability signifies better protection to the licensed users. In both cases the decision threshold plays a pivotal role in determining the fate of the unused spectrum. In this paper, we study the difficulty of selecting an appropriate threshold to minimize the error probability in an uncertain low SNR regime. The sensing failure issue is analyzed, and an optimal threshold is computed that yields minimum error rate. An adaptive double threshold concept has been proposed to make the detection robust and a closed-form equation for optimal threshold has been derived to minimize the error. The novel findings through simulation results exhibit improvement in Probability of detection and reduction in probability of error at low SNR in the presence of noise uncertainty factor.

Artificial intelligence has paved the way for different areas of computing such as speech recognition and translation, object detection, machine translation, and others. One of the goals of artificial general intelligence is to simulate human thinking and rationality within machines such that they are able to perceive their environment and then perform reasonable actions based on their perception. Creating a single model that performs every single task from visual perception to actuation is currently impossible. The system must be divided into several models each of which functions independently as well also contribute to the operation of the whole intelligent machine. In this paper, an intelligent sensing and caption generation (ISCG) system is proposed which is capable of detecting living/non-living objects and states of motion in images. The system consists of two separate modules of caption generator and intelligence engine with a Convolutional Neural Network (CNN) for determining the different objects in the images. Our model yields state-of-the-art performance on benchmarked dataset.

Traffic congestion is a major threat to transportation sector in every urban city around the world. This causes many adverse effects like, heavy fuel consumption, increased waiting time, pollution, etc. and pose an eminent challenge to the movement of emergency vehicles. To achieve better driving we proceed towards a trending research field called Social Internet of Vehicles (SIoV). A social network paradigm that permits the establishment of social relationships among every vehicle in the network or with any road infrastructure can be radically helpful. This holds as the aim of SIoV, to be beneficial for the drivers, in improving the road safety, avoiding mishaps, and have a friendly-driving environment. In this paper, we propose a Dynamic congestion control with Throughput Maximization scheme based on Social Aspect (D-TMSA) utilizing the social, behavioral and preference-based relationships. Our proposed scheme along with the various social relationship types allocates green signal to maximize the traffic flow passing through an intersection. Simulation results show that the D-TMSA outperforms the existing work by achieving high throughput, lowering the total traveling time and reducing the average waiting time to better the flow of traffic based on their social attributes with each other.

Over the last few years, deploying data to cloud service for repository is an appealing passion that avoids efforts on significant information sustenance and administration. In distributed repository utilities, deduplication technique is often exploited to minimize the capacity and bandwidth necesseties of amenities by erasing repetitive data and caching only a solitary duplicate of them. Proof-of-Ownership mechanisms authorize any possessor of the identical information to approve to the distributed repository server that he possess the information in a dynamic way. In repository utilities with enormous information, the repository servers may intend to minimize the capacity of cached information, and the customers may want to examine the integrity of their information with a reasonable cost. We propose Secure Deduplication and Virtual Auditing of Data in Cloud (SDVADC) mechanism that realizes integrity auditing and deduplication of information in cloud. The mechanism supports secure deduplication of information and effective virtual auditing of the documents during the download process. In addition, the proposed mechanism lowers the burden of dataowner to audit documents by himself and there is no need to delegate auditing to the Third Party Auditor (TPA). Experimental results demonstrate that the virtual auditing has low auditing time cost relative to the existing public auditing schemes.

Cloud repository is one of the most important services afforded by Cloud Computing where information is preserved, maintained, archived in distant servers and made available to the users over the Internet. Provided with the cloud repository facilities, customers can organize themselves as a cluster and distribute information with one another. In order to allow public integrity auditing on the information stored in semi-trusted cloud server, customers compute the signatures for every chunk of the shared information. When a malicious client is repudiated from the group, the chunks that were outsourced to the cloud server by this renounced customer need to be verified and re-signed by the customer present in the cluster (i.e., the straightforward approach) which results in huge transmission and reckoning cost for the customer. In order to minimize the burden of customers present in the cluster, in the existing scheme Panda, the semi-trusted Cloud Service Provider (CSP) is allowed to compute the Re−sign key. Further, the CSP audits and re-signs the revoked customer chunks by utilizing the Re−sign key. So, it is easy for the CSP by colluding with the revoked customer to find the secret keys of the existing customer. We introduce a novel Collusion Resistant User Revocable Public Auditing of Shared Data in Cloud (CRUPA) by making use of the concept of regression technique. In order to secure the secret keys of the existing customers from the CSP, we have allowed the information proprietor to compute the Re−sign key using the regression technique. Whenever the information proprietor revokes the customer from the cluster, the information proprietor computes the Re−sign key using the regression technique and sends to the CSP. Further, the CSP audits and re-signs the revoked customer chunks using the Re−sign key. The Re−sign key computed by the information proprietor using regression method is highly secure and the malicious CSP cannot find the private information of the customers in the cluster. Besides, our mechanism achieves significant improvement in the computation cost of the Re−sign key by information proprietor. Further, the proposed scheme is collusion resistant, supports effective and secure customer repudiation, multi-information proprietor batch auditing and is scalable.

Existing charging techniques for lithium-ion batteries use a largely open-loop approach, where the charge profile is predecided based on a priori knowledge of cell parameters. There is a need for closed-loop charging techniques that use instantaneous cell voltage and/or temperature to modulate the charging current magnitude. This paper addresses this gap by proposing a constant-temperature constant-voltage (CT-CV) charging technique, considering cell temperature as a key degradation metric. The proposed CT-CV charging scheme employs a simple and easy-to-implement proportional-integral-derivative (PID) controller aided by a feed-forward term. The charging current is dynamically adjusted in response to the battery temperature, which indirectly reflects its aging and thermal environment. As per experimental results, the proposed method achieves 20% faster charging with the same total temperature rise as constant-current constant-voltage (CC-CV) technique. Alternatively, it causes 20% lower cell temperature rise for given total charge time. It can easily accommodate applications that demand even faster charging by simply raising the set temperature. This paper establishes the benefits of the proposed CT-CV charging at cell level and raises the possibility of extending it to the pack level by integrating it with battery management systems.

Internet of Things (IoT) paradigm links physical objects in the real world to cyber world and enables the creation of smart environments and applications. A physical object is the fundamental building block of the IoT, known as a Smart Device , that can monitor the environment. These devices can communicate with each other and have data processing abilities. When deployed, smart devices collect real-time data and publish the gathered data on the Web. The functionality of smart devices can be abstracted as a service and an IoT application can be built by combining the smart devices with these services that help to address challenges of day-to-day activities. The IoT comprises billions of these intelligent communicating devices that generate enormous amount of data, and hence performing analysis on this data is a significant task. Using search techniques, the size and extent of data can be reduced and limited, so that an application can choose just the most important and valuable data items as per its necessities. It is, however, a tedious task to effectively seek and select a proper device and/or its data among a large number of available devices for a specific application. Search techniques are fundamental to IoT and poses various challenges like a large number of devices, dynamic availability, restrictions on resource utilization, real time data in various types and formats, past and historical monitoring. In the recent past, various methods and techniques have been developed by the research community to address these issues. In this paper, we present a review of the state-of-the-art search methods for the IoT, classifying them according to their design principle and search approaches as: IoT data and IoT object-based techniques. Under each classification, we describe the method adopted, their advantages and disadvantages. Finally, we identify and discuss key challenges and future research directions that will allow the next generation search techniques to recognize and respond to user queries and satisfy the information needs of users.

In today’s world, online shopping is very attractive and grown exponentially due to revolution in digitization. It is a crucial demand to provide recommendation for all the search engine to identify users’ need. In this paper, we have proposed a ANOVA Cosine Similarity Image Recommendation (ACSIR) framework for vertical image search where text and visual features are integrated to fill the semantic gap. Visual synonyms of each term are computed using ANOVA p value by considering image visual features on text-based search. Expanded queries are generated for user input query, and text-based search is performed to get the initial result set. Pair-wise image cosine similarity is computed for recommendation of images. Experiments are conducted on product images crawled from domain-specific site. Experiment results show that the ACSIR outperforms iLike method by providing more relevant products to the user input query.

The Internet of Things (IoT) paradigm envisions to support the creation of several applications that aids in the betterment of the society from various sectors such as environment, finance, industry etc. These applications are to be user-centric for their larger acceptance by the society. With the increase in the number of sensors that should are getting connected to the IoT infrastructure, there is an augmented increase in the amount of data generated by these sensors. Therefore it becomes a fundamental requirement to search for the sensors that produce the most applicable data required by the application. In this regard, context parameters of the sensors and the application users can be utilized to effectively filter out sensors from a large group. This paper proposes a sensor search scheme based on semantic-weights and fuzzy clustering. We have modified the traditional fuzzy c-means clustering algorithm by incorporating the semantic and context attributes of the sensors to obtain fuzzy clusters. During the query resolution phase, the query is directed to the most appropriate cluster. These clusters are formed through the use of linguistic variables rather than quantitative attributes and thus aid in effective user-centric search results. Experimental results indicate that the proposed scheme achieves better performance when compared to the existing techniques.

Search engines are an indispensable part of a web user's life. A vast majority of these web users experience difficulties caused by the keyword-based search engines such as inaccurate results for queries and irrelevant URLs even though the given keyword is present in them. Also, relevant URLs may be lost as they may have the synonym of the keyword and not the original one. This condition is known as the polysemy problem. To alleviate these problems, we propose an algorithm called automatic discovery and ranking of synonyms for search keywords in the web (ADRS). The proposed method generates a list of candidate synonyms for individual keywords by employing the relevance factor of the URLs associated with the synonyms. Then, ranking of these candidate synonyms is done using co-occurrence frequencies and various page count-based measures. One of the major advantages of our algorithm is that it is highly scalable which makes it applicable to online data on the dynamic, domain-independent and unstructured World Wide Web. The experimental results show that the best results are obtained using the proposed algorithm with WebJaccard.

Image recommendation is an important feature of search engine, as tremendous amount of images are available online. It is necessary to retrieve relevant images to meet the user’s requirement. In this paper, we present an algorithm image recommendation with absorbing Markov chain (IRAbMC) to retrieve relevant images for a user’s input query. Images are ranked by calculating keyword relevance probability between annotated keywords from log and keywords of user input query. Keyword relevance is computed using absorbing Markov chain. Images are reranked using image visual features. Experimental results show that the IRAbMC algorithm outperforms Markovian semantic indexing (MSI) method with improved relevance score of retrieved ranked images.

Cognitive Radio based Wireless Sensor Network is a novel concept that integrates the dynamic spectrum access capability of cognitive radio into wireless sensor networks for the futuristic sensor networks and wireless communication technology. Spectrum sensing plays a quintessential role in a cognitive radio network but is a major constraint for a battery powered sensor with stringent energy limitations. The spectrum sensing algorithms are expected to yield acceptable detection probability at low SNR under noise uncertainty with minimum power consumption in a WSN. In this paper, a new spectrum sensing method has been proposed to overcome sensing failure under low SNR environment. The proposed technique is based on adaptive double threshold theory which improves the detection performance by 39.63 and 27.22% at SNR = −10dB as compared to the conventional energy detection and available double threshold-based method respectively. Furthermore, the proposed method of spectrum sensing is evaluated for its deployment into a CR-WSN using the evaluation metrics: Time and Sample Complexity. The comparative evaluation of the spectrum sensing method in a WSN through simulations shows that the proposed technique offers substantial reduction in sample and time complexity of the wireless sensor nodes.

The Jansen mechanism is a one degree-of-freedom, planar, 12-link, leg mechanism that can be used in mobile robotic applications and in gait analysis. This paper presents the kinematics and dynamics of the Jansen leg mechanism. The forward kinematics, accomplished using circle intersection method, determines the trajectories of various points on the mechanism in the chassis (stationary link) reference frame. From the foot point trajectory, the step length is shown to vary linearly while step height varies non-linearly with change in crank radius. A dynamic model for the Jansen leg mechanism is proposed using bond graph approach with modulated multiport transformers. For given ground reaction force pattern and crank angular speed, this model helps determine the motor torque profile as well as the link and joint stresses. The model can therefore be used to rate the actuator torque and in design of the hardware and controller for such a system. The kinematics of the mechanism can also be obtained from this dynamic model. The proposed model is thus a useful tool for analysis and design of systems based on the Jansen leg mechanism. (C) 2015 Elsevier B.V. All rights reserved.

Cloud computing provides economical and effective solution for sharing data among cloud users with low maintenance cost. The security of data and identity confidentiality while sharing data in multi-owner way cannot be assured by the Cloud Service Providers (CSP's). The Cloud Service Providers are reliable but curious to know the recurrent membership changes in the cloud. In this paper, we propose a secure multi-owner data sharing for dynamic group in the cloud with RSA Chinese Remainder Theorem (RSA-CRT) encryption technique and substring index generation method. RSA-CRT efficiently manages revocation list, key management, with reduced storage and computational overhead. The substring Index generation algorithm reduces the storage space compared to wild card fuzzy alogorithm(1). (C) 2016 The Authors. Published by Elsevier B.V.

The availability of digital documents over web has increased exponentially and hence there is a need for effective methods to retrieve and organise. Since data is dispersed globally and unorganised...

As internet of things (IoT) is overpopulated with a multitude of objects, services and interactions locating the most relevant object is emerging as a major obstacle. Over the last few years, the social internet of things (SIoT) paradigm, where objects independently establish social relationships with the other things has become popular as it provides several new characteristics to carryout reliable discovery approaches. Given a large scale deployment of socially connected objects, finding the shortest path to reach the service provider remains as a fundamental challenge. Most of the existing techniques, search for a specific object or service utilising its friendship or friends of friends connections. As a result, each object has to manage a large set of friends, thus slowing down the search process. In this paper, we propose a similarity based object search mechanism that dynamically creates and manages relationships based on physical location proximity and social context of users in social communities. The result shows enhancement in the proposed method over the existing search techniques FSS, FSASV and LSFGA in terms of local cluster coefficients, the average degree of connections and average path length.

The Digital Image processing applications like medical imaging, satellite imaging, Biometric trait images etc., rely on multipliers to improve the quality of image. However, existing multiplication techniques introduce errors in the output with consumption of more time, hence error free high speed multipliers has to be designed. In this paper we propose FPGA based Recursive Error Free Mitchell Log Multiplier (REFMLM) for image Filters. The 2x2 error free Mitchell log multiplier is designed with zero error by introducing error correction term is used in higher order Karastuba-Ofman Multiplier (KOM) Architectures. The higher order KOM multipliers is decomposed into number of lower order multipliers using radix 2 till basic multiplier block of order 2x2 which is designed by error free Mitchell log multiplier. The 8x8 REFMLM is tested for Gaussian filter to remove noise in fingerprint image. The Multiplier is synthesized using Spartan 3 FPGA family device XC3S1500-5fg320. It is observed that the performance parameters such as area utilization, speed, error and PSNR are better in the case of proposed architecture compared to existing architectures

This paper presents, a Group Based Optimal Retransmission Medium Access (GORMA) Protocol is designed that combines protocol of Collision Avoidance (CA) and energy management for low-cost, short-range, low-data rate and low-energy sensor nodes applications in environment monitoring, agriculture, industrial plants etc. In this paper, the GORMA protocol focuses on efficient MAC protocol to provide autonomous Quality of Service (QoS) to the sensor nodes in one-hop QoS retransmission group and two QoS groups in WSNs where the source nodes do not have receiver circuits. Hence, they can only transmit data to a sink node, but cannot receive acknowledgement control signals from the sink node. The proposed protocol GORMA provides QoS to the nodes which work independently on predefined time by allowing them to transmit each packet an optimal number of times within a given period. Our simulation results shows that the performance of GORMA protocol, which maximize the delivery probability of one-hop QoS group and two QoS groups and minimize the energy consumption., Comment: 9 pages in IEEE format and 6 figures

Many of the adaptive watermarking schemes reported in the literature consider only local audio signal properties. Many schemes require complex computation along with manual parameter settings. In this paper, we propose a novel, fuzzy, adaptive audio watermarking algorithm based on both global and local audio signal properties. The algorithm performs well for dynamic range of audio signals without requiring manual initial parameter selection. Here, mean value of energy (MVE) and variance of spectral flux (VSF) of a given audio signal constitutes global components, while the energy of each audio frame acts as local component. The Quantization Index Modulation (QIM) step size Δ is made adaptive to both the global and local features. The global component automates the initial selection of Δ using the fuzzy inference system while the local component controls the variation in it based on the energy of individual audio frame. Hence Δ adaptively controls the strength of watermark to meet both the robustness and inaudibility requirements, making the system independent of audio nature. Experimental results reveal that our adaptive scheme outperforms other fixed step sized QIM schemes and adaptive schemes and is highly robust against general attacks.

Energy management is one of the critical parameters in Wireless Sensor Networks. In this paper we attempt for a solution to balance the energy usage for maximizing the network lifetime, increase the packet delivery ratio and throughput. Our proposed algorithm is based on Energy Density of the clusters in Wireless Sensor Networks. The cluster head is selected using two step method and on-demand routing approach to calculate the balanced energy shortest path from source to sink. This unique approach maintains the balanced energy utilization among all nodes by selecting the different cluster heads dynamically. Our simulation results have compared with one of the plain routing scheme (EBRP) and cluster based routing (TSCHS), which shows the significant improvements in minimizing the delay and energy utilization and maximizing the network lifetime and throughput with respect to these works., Comment: 18 pages,7 Figures

Similarity matching and join of time series data streams has gained a lot of relevance in today's world that has large streaming data. This process finds wide scale application in the areas of location tracking, sensor networks, object positioning and monitoring to name a few. However, as the size of the data stream increases, the cost involved to retain all the data in order to aid the process of similarity matching also increases. We develop a novel framework to addresses the following objectives. Firstly, Dimension reduction is performed in the preprocessing stage, where large stream data is segmented and reduced into a compact representation such that it retains all the crucial information by a technique called Multi-level Segment Means (MSM). This reduces the space complexity associated with the storage of large time-series data streams. Secondly, it incorporates effective Similarity Matching technique to analyze if the new data objects are symmetric to the existing data stream. And finally, the Pruning Technique that filters out the pseudo data object pairs and join only the relevant pairs. The computational cost for MSM is O(l*ni) and the cost for pruning is O(DRF*wsize*d), where DRF is the Dimension Reduction Factor. We have performed exhaustive experimental trials to show that the proposed framework is both efficient and competent in comparison with earlier works., Comment: 20 pages,8 figures, 6 Tables

Web sequential patterns are important for analyzing and understanding users behaviour to improve the quality of service offered by the World Wide Web. Web Prefetching is one such technique that utilizes prefetching rules derived through Cyclic Model Analysis of the mined Web sequential patterns. The more accurate the prediction and more satisfying the results of prefetching if we use a highly efficient and scalable mining technique such as the Bidirectional Growth based Directed Acyclic Graph. In this paper, we propose a novel algorithm called Bidirectional Growth based mining Cyclic behavior Analysis of web sequential Patterns (BGCAP) that effectively combines these strategies to generate prefetching rules in the form of 2-sequence patterns with Periodicity and threshold of Cyclic Behaviour that can be utilized to effectively prefetch Web pages, thus reducing the users perceived latency. As BGCAP is based on Bidirectional pattern growth, it performs only (log n+1) levels of recursion for mining n Web sequential patterns. Our experimental results show that prefetching rules generated using BGCAP is 5-10 percent faster for different data sizes and 10-15% faster for a fixed data size than TD-Mine. In addition, BGCAP generates about 5-15 percent more prefetching rules than TD-Mine., Comment: 19 pages

Wireless Sensor Networks are susceptible to communication failures and security attacks due to broadcast nature of communication. Multipath based communication techniques were designed to address communication failure. This paper proposes Secure Approximate Data Aggregation (SADA) in which synopsis are generated using primitive polynomial and Message Authentication Codes (MACs) are transmitted along with the synopsis to ensure integrity. SADA provides data freshness and integrity at a communication cost of O(1). Simulation results show that the SADA protocol incurs lower energy consumption and communication and computation cost compared to the state-of-the art protocols [1] [2].

One of the most fundamental services of cloud computing is Cloud storage service. Huge amount of sensitive data is stored in the cloud for easy remote access and to reduce the cost of storage. The confidential data is encrypt before uploading to the cloud server in order to maintain privacy and security. All conventional searchable symmetric encryption(SSE) schemes enable the users to search on the entire index file. In this paper, we propose the Domain and Range Specific Index Generation(DRSIG) scheme that minimizes the Index Generation time. This scheme adopts collection sort technique to split the index file into D Domains and R Ranges. The Domain is based on the length of the keyword; the Range splits within the domain based on the first letter of the keyword. A mathematical model is used to encrypt the indexed keyword that eliminates the information leakage. The time complexity of the index generation is O(NT × 3) where NT - Number of rows in index document and 3 is Number of columns in index document. Experiments have been conducted on real world dataset to validate proposed DRSIG scheme. It is observed that DRSIG scheme is efficient and provide more secure data than Ranked Searchable Symmetric Encryption(RSSE) Scheme.

In this paper, we propose a high capacity, self-synchronized, lossless audio watermarking algorithm based on the alpha (‘ �’) statistic modulation. Here ‘ �’ is related to the correlation among any given sequence i.e audio samples and it is modulated according to the watermark bit stream. The embedding scheme is tested in both the time domain and DWT domain. Though the time domain embedding reduces the computational time in searching the synchronization codes, the time-frequency localization capability of DWT provides good trade off between the computational complexity and robustness of synchronization codes. In case of DWT, ‘�’ related to the 2 nd level DWT coarse wavelet components is used for embedding the watermark. The offset value used for embedding is made adaptive to the required SNR for the final watermarked audio signal. After extraction of the embedded watermark using a watermark key, original audio can be recovered with minimal distortion. The watermarking method presented here does not require the use of the original signal for watermark detection. Also high embedding capacity is achieved by using small sized audio frames. Experimental results reveal that the proposed watermarking scheme maintains high audio quality and is simultaneously highly robust to pirate attacks, including MP3 compression, cropping, filtering, re-sampling, and re-quantization.

Real-time industrial application requires routing protocol that guarantees data delivery with reliable, efficient and low end-to-end delay. Existing Routing(THVR) [13] is based velocity of Two-Hop Velocity and protocol relates two-hop velocity to delay to select the next forwarding node, that has overhead of exchanging control packets, and depleting the available energy in nodes. We propose a Real-Time Reliable Data delivery based on Virtual Coordinates Routing (RRDVCR) algorithm, based on the number of hops to the destination rather than geographic distance. Selection of forwarding node is based on packet progress offered by two-hops, link quality and available energy at the forwarding nodes. All these metric are co-related by dynamic co-relation factor. The proposed protocol uses selective acknowledgment scheme that results in lower overhead and energy consumption. Simulation results shows that there is about 22% and 9.5% decrease in energy consumption compared to SPEED [8] and THVR [13] respectively, 16% and 38% increase in packet delivery compared to THVR [13] and SPEED [8] respectively, and overhead is reduced by 50%.

Data mining involves nontrivial process of extracting knowledge or patterns from large databases. Genetic Algorithms are efficient and robust searching and optimization methods that are used in data mining. In this paper we propose a Self-Adaptive Migration Model GA (SAMGA), where parameters of population size, the number of points of crossover and mutation rate for each population are adaptively fixed. Further, the migration of individuals between populations is decided dynamically. This paper gives a mathematical schema analysis of the method stating and showing that the algorithm exploits previously discovered knowledge for a more focused and concentrated search of heuristically high yielding regions while simultaneously performing a highly explorative search on the other regions of the search space. The effective performance of the algorithm is then shown using standard testbed functions and a set of actual classification datamining problems. Michigan style of classifier was used to build the classifier and the system was tested with machine learning databases of Pima Indian Diabetes database, Wisconsin Breast Cancer database and few others. The performance of our algorithm is better than others.

Multimedia mining primarily involves, information analysis and retrieval based on implicit knowledge. The ever increasing digital image databases on the Internet has created a need for using multimedia mining on these databases for effective and efficient retrieval of images. Contents of an image can be expressed in different features such as Shape, Texture and Intensity-distribution(STI). Content Based Image Retrieval(CBIR) is an efficient retrieval of relevant images from large databases based on features extracted from the image. Most of the existing systems either concentrate on a single representation of all features or linear combination of these features. The paper proposes a CBIR System named STIRF (Shape, Texture, Intensity-distribution with Relevance Feedback) that uses a neural network for nonlinear combination of the heterogenous STI features. Further the system is self-adaptable to different applications and users based upon relevance feedback. Prior to retrieval of relevant images, each feature is first clustered independent of the other in its own space and this helps in matching of similar images. Testing the system on a database of images with varied contents and intensive backgrounds showed good results with most relevant images being retrieved for a image query. The system showed better and more robust performance compared to existing CBIR systems.

This paper deals with the design and implementation of an image coding algorithm based on fixed threshold method. Threshold is the Peak Absolute Error (PAE) allowed in the reconstructed image. It has been shown that lossless edges with near-lossless/lossless filled area give a high fidelity images. Results are compared with Set Partitioning In Hierarchical Tree (SPIHT) [A. Said, W.A. Pearlman, A new, fast, and efficient image codec based on set partitioning in hierarchical trees, IEEE Trans. Circ. Syst. Video Technol. 12 (1996) 243–250] method. Experimented results show that the proposed method provides high fidelity images, and it has been shown that these images are visually better than images reconstructed using SPIHT algorithm for the same compression ratio. The execution time of the algorithm is almost half compared to that of the SPIHT algorithm which requires wavelet transform of an image.

In this paper, we propose a novel method using wavelets as input to neural network self-organizing maps and support vector machine for classification of magnetic resonance (MR) images of the human brain. The proposed method classifies MR brain images as either normal or abnormal. We have tested the proposed approach using a dataset of 52 MR brain images. Good classification percentage of more than 94% was achieved using the neural network self-organizing maps (SOM) and 98% from support vector machine. We observed that the classification rate is high for a support vector machine classifier compared to self-organizing map-based approach.

In this paper we propose a novel method for designing block-wise lossless image compression scheme using linear predictors. In this prediction scheme, the prediction for each pixel is formed by using a set of least-square-based linear prediction coefficients of the block to which the current pixel belongs. Predicted value of each pixel is subtracted from the actual value of the current pixel to get an error image. An error image is compressed using grayscale bit plane coding using quadtree method. Experimental results show that the compression performance of the proposed method is superior to Joint Photographics Expert Group's [1] JPEG-LS [IEEE Trans. Image Processing 9 (2000) 1309] method, and Classified Adaptive Prediction and Entropy Coding in terms of coding performance.

This paper presents the parallel implementation of the neocognitron neural network paradigm on star topology, A detailed theoretical timing analysis has been presented along with a comparison between the experimental and theoretical results. These studies demonstrate that a linear speedup can be achieved by mapping the neocognitron onto a star topology.

Genetic programming (GP) is an evolutionary technique and is gaining attention for its ability to learn the underlying data relationships and express them in a mathematical manner. Although GP uses the same principles as genetic algorithms, it is a symbolic approach to program induction; i.e., it involves the discovery of a highly fit computer program from the space of computer programs that produces a desired output when presented with a particular input. We have successfully applied the GP paradigm for the n-category pattern classification problem. The ability of the GP classifier to learn the data distributions depends upon the number of classes and the spatial spread of data. As the number of classes increases, it increases the difficulty for the GP classifier to resolve between classes. So, there is a need to partition the feature space and identify sub-spaces with reduced number of classes. The basic objective is to divide the feature space into sub-spaces and hence the data set that contains representative samples of n classes into sub-data sets corresponding to the sub-spaces of the feature space, so that some of the sub-data sets/spaces can have data belonging to only p-classes (p

This paper explores the feasibility of applying genetic programming (GP) to multicategory pattern classification problem for the first time. GP can discover relationships among observed data and express them mathematically. Multicategory pattern classification has been done traditionally by using the maximum likelihood classifier (MLC). GP-based techniques have an advantage over statistical methods because they are distribution free, i.e., no prior knowledge is needed about the statistical distribution of the data. GP also has the ability to automatically discover the discriminant features for a class. GP has been applied for two-category (class) pattern classification, In this paper, a methodology for GP-based n- class pattern classification is developed. The given n-class problem is modeled as n two-class problems, and a genetic programming classifier expression (GPCE) is evolved as a discriminant function for each class, The GPCE is trained to recognize samples belonging to its own class and reject samples belonging to other classes. A strength of association (SA) measure is computed for each GPCE to indicate the degree to which it can recognize samples belonging to its own class. The higher the value of SA, the better is the ability of a GPCE to recognize samples belonging to its own class and reject samples belonging to other classes. The SA measure is used for uniquely assigning a class to an input Feature vector. Heuristic rules are used to prevent a GPCE with a higher SA from swamping a GPCE with a lower SA, Experimental results are presented to demonstrate the applicability of CP for multicategory pattern classification, and the results obtained are found to be satisfactory, and are compared with those of the MLC, We also discuss the various issues that arise in our approach to GP- based classification, such as the creation of training sets, the role of incremental learning, and the choice of function set in the evolution of GPCEs, as well as conflict resolution for uniquely assigning a class.

In recent years, parallel computers have been attracting attention for simulating artificial neural networks (ANN). This is due to the inherent parallelism in ANN. This work is aimed at studying ways of parallelizing adaptive resonance theory (ART), a popular neural network algorithm. The core computations of ART are separated and different strategies of parallelizing ART are discussed. We present mapping strategies for ART 2-A neural network onto ring and mesh architectures. The required parallel architecture is simulated using a parallel architectural simulator, PROTEUS and parallel programs are written using a superset of C for the algorithms presented. A simulation-based scalability study of the algorithm-architecture match is carried out. The various overheads are identified in order to suggest ways of improving the performance. Our main objective is to find out the performance of the ART2-A network on different parallel architectures. (C) 1999 Elsevier Science B.V. All rights reserved.

In recent years, wireless communication technology has reduced the distance between people and has hence become a significant part of our lives. Two such technologies are WiFi(IEEE 802.11) and WiMAX(IEEE 802.16) where the latter is a long range system covering many kilometers, whereas former is a synonym for WLAN providing a coverage of only short ranges. This work describes the implementation of a framework in which a multi-hop, ad-hoc network is deployed with hybrid nodes to enhance network throughput. The data traffic received is split between the WiFi and WiMAX radios on the basis of th e split coefficient value statically. The routing algorithm being implemented in this paper is the be e-hive algorithm. Bee-hive algorithm is a multi-path routing algorithm inspired by the social behavior of swarms of bees. It is dynamic, robust and flexible yet simple algorithm which can prove helpful for optimal management of available network resources. In this paper, we have split data traffic over two radio channels for achieving enhanced performance and reduced delay.

Opportunistic Routing in wireless sensor networks is a multi-hop routing. In this routing neighbors of a node overhear the transmission and form multiple hops from source to the destination for transfer of information. The set of neighbor nodes participating in the routing are included in the forwarder list in the order of priority. A node with highest priority is allowed to forward the packet it hears. This paper implements Energy Efficient Selective Opportunistic Routing (EESOR), reduces the size of forwarder list by applying a condition that the forwarding node is nearer to the destination. The path followed by acknowledgment packet follows opportunistic routing, assuring reliability of transmission and energy balancing. The simulated results obtained in NS2 simulator show that proposed EESOR protocol performs better than existing Energy Efficient Opportunistic Routing (EEOR) protocol in terms of average End-to-End delay, maximum End-to-End delay and Network Lifetime.

Genetic algorithms (GAs) are search methods that are being employed in a multitude of applications with extremely large search spaces. Recently, there have been considerable interest among GA researchers in understanding and formalizing the working of GAs. In an earlier paper, we have introduced the notion of binomially distributed populations as the central idea behind an exact “populationary” model of the large-population dynamics of the GA operators for objective functions called “functions of unitation.” In this paper, we extend this populationary model of GA dynamics to a more general class of objective functions called functions of unitation variables . We generalize the notion of a binomially distributed population to a generalized binomially distributed population (GBDP). We show that the effects of selection, crossover, and mutation can be exactly modelled after decomposing the population into GBDPs. Based on this generalized model, we have implemented a GA simulator for functions of two unitation variables—GASIM 2, and the distributions predicted by GASIM 2 match with those obtained from actual GA runs. The generalized populationary model of GA dynamics not only presents a novel and natural way of interpreting the workings of GAs with large populations, but it also provides for an efficient implementation of the model as a GA simulator.

The Artificial Neural Networks (ANNs) are being used to solve a variety of problems in pattern recognition, robotic control, VLSI CAD and other areas. In most of these applications, a speedy response from the ANNs is imperative. However, ANNs comprise a large number of artificial neurons, and a massive interconnection network among them. Hence, implementation of these ANNs involves execution of computer-intensive operations. The usage of multiprocessor systems therefore becomes necessary. In this article, we have presented the implementation of ART1 and ART2 ANNs on ring and mesh architectures. The overall system design and implementation aspects are presented. The performance of the algorithm on ring, 2-dimensional mesh and n-dimensional mesh topologies is presented. The parallel algorithm presented for implementation of ART1 is not specific to any particular architecture. The parallel algorithm for ARTE is more suitable for a ring architecture.

We present through the use of Petri Nets, modeling techniques for digital systems realizable using FPGAs. These Petri Net models are used for logic validation at the logic design phase. The technique is illustrated by modeling practical circuits. Further, the utility of the technique with respect to timing analysis of the modeled digital systems is considered.

This paper introduces a formalism to model task execution in distributed computing systems. The proposed formalism, based on a partial order semantic model, extends standard temporal logic to specify task execution in distributed systems. It allows derivation of a set of inequalities sufficient to ensure deadline constraints. Task execution in an onboard computer is modeled and verified using the developed formalism. Extensions for task refinement are also presented.

We describe the design of a directory-based shared memory architecture on a hierarchical network of hypercubes. The distributed directory scheme comprises two separate hierarchical networks for handling cache requests and transfers. Further, the scheme assumes a single address space and each processing element views the entire network as contiguous memory space. The size of individual directories stored at each node of the network remains constant throughout the network. Although the size of the directory increases with the network size, the architecture is scalable. The results of the analytical studies demonstrate superior performance characteristics of our scheme compared with those of other schemes.

In this paper a pipelined ring algorithm is presented for efficient computation of one and two dimensional Fast Fourier Transform (FFT) on a message passing multiprocessor. The algorithm has been implemented on a transputer based system and experiments reveal that the algorithm is very efficient. A model for analysing the performance of the algorithm is developed from its computation-communication characteristics. Expressions for execution time, speedup and efficiency are obtained and these expressions are validated with experimental results obtained on a four transputer system. The analytical model is then used to estimate the performance of the algorithm for different number of processors, and for different sizes of the input data.

In executing tasks involving intelligent information processing, the human brain performs better than the digital computer. The human brain derives its power from a large number [O(1011)] of neurons which are interconnected by a dense interconnection network [O(105) connections per neuron]. Artificial neural network (ANN) paradigms adopt the structure of the brain to try to emulate the intelligent information processing methods of the brain. ANN techniques are being employed to solve problems in areas such as pattern recognition, and robotic processing. Simulation of ANNs involves implementation of large number of neurons and a massive interconnection network. In this paper, we discuss various simulation models of ANNs and their implementation on distributed memory systems. Our investigations reveal that communication-efficient networks of distributed memory systems perform better than other topologies in implementing ANNs.

A new fault-tolerant multi-transputer architecture capable of tolerating failure of any one component in the system is described. In the proposed architecture the processing nodes are automatically reconfigured in the event of a fault and the computations continue from the stage where the fault occurred. The process of reconfiguration is transparent to the user, and the identity of the failed component is communicated to the user along with the results of computations. Parallel solution of a typical engineering problem involving solution of Laplace's equation by the boundary element method has been implemented. The performance of the architecture in the event of faults has been investigated.

This paper is concerned with the movable-separability properties of isothetic simple polygons. Previously obtained results in this area applied only to the highly restricted class of star-shaped isothetic polygons. We introduce new classes of isothetic polygons which are substantially more general, and establish their movable-separability properties. We also define new structures called separability hulls of isothetic polygons, and introduce the notion of a jagged cross , which is a very broad subclass of isothetic simple polygons with highly interesting geometry. Elegant results on the separability-related properties of these structures are shown. Although occasional remarks on algorithmic aspects are made, the emphasis is more on structural geometric properties of isothetic polygons than on algorithmic implications, which are open for further investigation.

Mahadevan, I. and L.M. Patnaik, Performance evaluation of bidirectional associative memory on a transputer-based parallel system, Parallel Computing 18 (1992) 401-413. In this paper we discdss a parallel implementation of an artificial neural network for pattern association, the Bidirectional Associative Memory (BAM). The parallelism inherent in BAM has been exploited using various transputer-based architectures like hypercube, mesh and linear array. The speedup and utilization of the system are obtained for the architectures mentioned above and a comparative study is made. Simulation results show that the hypercube topology gives a better performance in terms of speedup and utilization when compared to other architectures.

A new interconnection topology-the extended hypercube-consisting of an interconnection network of k-cubes is discussed. The extended hypercube is a hierarchical, expansive, recursive structure with a constant predefined building block. The extended hypercube retains the positive features of the k-cube at different levels of hierarchy and at the same time has some additional advantages like reduced diameter and constant degree of a node. The paper presents an introduction to the topology of the extended hypercube and analyzes its architectural potential in terms of message routing and executing a class of highly parallel algorithms. Topological properties and performance studies of the extended hypercube are presented. >

The way in which Wireless Sensor Networks (WSN) are designed requires that energy be taken as the most crucial element if WSNs are to be used in the most effective way to serve the purpose for which they have been deployed in the target region. Routing protocols are the main aids that can assist in reducing the energy consumption required by the transmission of data throughout the sensor networks. In this paper we have proposed a flat routing protocol for sensor networks. This approach is one of the simplest protocols in terms of the route determination process and the number of messages through the network. Routes from the source to the destination are determined by use of only the hop count and remaining energy of the neighbor nodes for each node. The proposed protocol is simulated and is compared with the flood routing protocol.

As security is a major concern in present times, reliable authentication systems are in great demand. A biometric trait like the vascular pattern on the back of the hand of a person is unique and secure. A biometric system working on this principle often fails to authenticate a person either because of the varying hand posture or due to an injury altering the vein pattern. In this paper we propose an authentication system to overcome these disadvantages by modularizing the image and then comparing the features. This method of authentication reduces the False Rejection Ratio (FRR) and also False Acceptance Ratio (FAR) of the system.