Computer Engineering

Visual cryptography (VC) is a technique for protecting the secret image which encodes the picture into many shares and allots them to different members. At the point when all shares are adjusted and stacked together, they uncover the secrete image. In k out of n (k, n) VC plot, the secret image is shared into n shares to such an extent that when k or more members by storing up their transparencies by methods for an overhead projector to uncovers the secret image. This paper proposed a spic and span of basic and strong (k, n) visual cryptography method which is utilized to successfully imparting the secret image to most extreme classification. In share creation process, determined new condition for arbitrary grids and after that XOR activities are performed to produce the 'n' transparencies. It is conceivable to decipher the secret image outwardly by superimposing a k subset of transparencies. All things considered no mystery information can be procured from the superposition of an unlawful subset. Trials, measurable and security evaluations are completed on the offers to approve the quality of the proposed plot by methods for a succession of examinations, for example, visual testing, encryption quality testing, security investigation and diverse assaults. The proposed (k, n) VC plot offers a reliable insurance for conveying pictures over people in general channels.

Cloud delivers computing resources as a service. Finest usage of cloud services is used to store information in the cloud servers. Pool of virtual servers is configured to store the users’ data with low cost. Cloud offers more advantages to the cloud users. Apart from the advantages and benefits of cloud, it has more security issues and vulnerability on the data stored in the cloud. This paper proposes asecurity service named AROMONcrypt to address the security issues in cloud storage. This security service is used to secure the data in cloud storage. AROMONcrypt uses two types of techniques to ensure the confidentiality of data namely encryption and obfuscation. This paper also describes Security as a Service (SEaaS). SEaaS provides AROMONcrypt security service from the Cloud Service Provider (CSP). Simulation is conducted with AROcrypt and MONcrypt security services considering time and security level. AROMONcrypt provides optimum security, and MONcrypt has taken minimum time.

Biometric technology has become a basis of an extensive array of highly secure identification and personal verification solutions in our world today. More importantly in the wake of heightened concern about security and rapid advancements in communication and mobility. Significant application areas of biometric systems include security monitoring, access control and authentication, border control and immigration, forensic investigation, telemedicine and so on. When a single trait is used in an application it is referred to as unimodal biometric, while combination of two or more sources or traits in an application is referred to as multiple biometrics. But biometric system that uses a single biometric trait for recognition has this propensity to contend with problems related to non-universality of the trait, spoof attacks, large intra-class variability, and noisy data. Besides, no single biometric trait can meet all the requirements of every possible application, hence the need for multiple biometric system to overcome the limitation of unimodal biometric. The new paradigm is robust against individual sensor or subsystem failures and spoof attack, as it is very difficult to spoof multiple traits simultaneously. In addition, the technological environment is very appropriate because of the widespread deployment of multimodal devices (PDAs, 3G mobile phones, Tablet PCs, laptops etc). The aim of this paper is to present an overview of multiple biometric systems, design approach and application scenario.

Collaborative filtering is one of the most widely used techniques in the recommender systems. determining the similarity between users is the base of collaborative filtering based recommender systems. obviously, choosing a proper similarity function improves the accuracy of recommendation in recommender systems. In this paper we provided a new similarity measure using features and have compared results with improved similarity measure and pearson traditional similarity measure. the comparison results show that our proposed method not only increases the accuracy of recommender, but also increases the prediction quality. .

In today’s world technology is getting more advanced, we have new technology for our personal living, even at home. Home automation is becoming more popular around the world and is becoming a common practice. The main concept of home automation is to automate everything in the house which can be done using technology to control and do the tasks that we would do manually. In this paper, we illustrate use of remote devices such as mobile phone, tablet or desktop and laptops to control, monitor the Home Appliances. In today's era, technology can enhance comfort zone of human life. Technology is evolving rapidly. By using the latest technology for home automation, we can build a fully automated home. By Using Raspberry Pi and Aurdino, a home automation system is built and is capable of operating home appliances automatically. In this paper we have implemented home automation for controlling electrical home appliances. We have provided the facility to control these devices through web as well as through mobile. Here we are using Raspberry Pi-3 as a server and Arduino to get the signals and send the same to the server. This paper proposes a very economical system for home automation.

Spam is the sending unwanted e-mail messages frequently with different contents, in large quantities to an indiscriminate set of recipients, and often proselytes a service or a website. Many intelligent systems have been developed for detecting spam emails, but many of them don’t have enough speed. In this paper, a fuzzy spam detection system in text classification mode is described that has been implemented in MATLAB. Because of the ANFIS uses the approximation capability of FIS and ANN as adaptive, it acts simple and powerful. In the proposed method, first extractor starts to extract all the tokens in the body of all emails. Genetic algorithm (GA) is then applied, to select the appropriate features of the tokens. These features are saved in a dictionary. Then ANFIS uses this dictionary for classifying emails. In this project, ANFIS has three inputs and one output. For obtaining ANFIS’ inputs, calculate a spamicity for each token. This criterion shows the rate of dangerous of each token. Then tokens of each email are classified into three categories, based on the amounts of their spamicity. Counts of tokens in each category, are three inputs to ANFIS system. ANFIS’ output determines that each email is spam or not.

Communities with different profiles shows the interests of community members. User and check-in venue details are used to cluster. Multimode multi-attribute edge-centric coclustering model is help to discover overlapping the communities. Overlapping communities is used to repair by replacing each edge with its vertices in edge clusters. Intermode and intra mode features are helped to us for making the process. Three intra mode features are used in the community detection process. M2 Clustering algorithm is used for community detection it is Edge clustering based on k-means and HM2 Clustering algorithm is used to detect overlapping communities of LBSNs and also called as Two-step hierarchical edge clustering. The overlapping community detection mechanism is enhanced with recommendation models. LBSNs, analyzing 1) the data source used, 2) the methodology community. 3) the objective of the community. The community are also classified with location. The system is enhanced with feature selection and feature fusion mechanism. This system also indicate the cluster of the two communities.

We are living in a very congested world where there are more and more vehicles. The noise and interference produced by these vehicles are enormous and their unwanted amplification of noise causes irritation. We aim to present a blind technique, enhancing speech, attenuating any kind of noise, by designing an adaptive filter depending on the nature of noise. The differentiation between speech and noise is made. This characteristic is used to derive a cost functional for speech enhancement. Adaptation is by changing two sets of weights. The simulation and synthesis of the above technique/algorithm is implemented using Verilogger pro and Leonardo spectrum.

The paper presents a system developed for recognizing gestures of Indian sign language from images of gestures. The proposed system is based on Discrete Cosine Transform (DCT) and neural networks used for gesture pattern recognition. Unlike the systems proposed by other researchers such as using a radio frequency or colored gloves to achieve the recognition our system does not impose any such constraints. Features are extracted from the images using DCT which greatly reduces the size of the feature vector. Neural networks error back propagation algorithm is used to recognize gestures of alphabets of English language. The system was implemented with 130 sample images of gestures of alphanumeric characters with a maximum of 5 images per gesture. Experimental results show that the neural network is able to recognize gestures with an accuracy of 99.52%.

For every human being there is no life without breathing. Respiration plays a vital role in all living beings. This respiration get affected by many factors. Particularly lung tumor causes major problem during inspiration and expiration. So it is necessary to develop a technique which gives detailed knowledge about the breathing dynamics. The techniques to detect lung tumours in process are invasive in some way whereas 3D modelling of lung will be helpful in detecting lung tumors which is a non-invasive technique. Thus, computer aided modelling of respiratory motion gains importance. In this paper we present an approach to model thermal expansion sensor using MEMS which helps to detect lung tumors in miniatureway.