Computer Engineering

Regression testing leads to running many tests many times. Hence it requires more cost. The most robust and straight forward technique for regression testing is to accumulate all integration tests and rerun them whenever new components are integrated into the system. This requires developers to keep all tests up-to-date, to evolve them as the subsystem interfaces changes and to add new integration tests as new services or new subsystems are added. As regression testing can become time consuming, test suite minimization (also known as Test Suite Reduction) technique is best suited to tackle it . In this paper we have explained the heuristic approach to solve this optimization problem.

This paper presents a Product recommendation system based on metric analysis of product descriptions. The developed system ranks the catalog of products and offers corresponding items to the user’s request while, at the same time, selecting the most diverse items. An algorithm for ranking is developed. Based on the request, the recommendation system finds the distance from this request to all documents from the collection of data. The request and the collection of data are sets of features. The system ranks the results in accordance with the following rules: minimizes the distance from the query to the relevant results, maximizes the distance from the query to the irrelevant results and maximizes the distance between the relevant query results. For ranking, Heterogeneous Euclidean-Overlap Metric (HEOM) of clothes catalogue items is used. HEOM metric uses different attribute distance functions to measure distances between objects in mixed scales. A dataset of clothes catalogue items is collected. The system, in addition to the basic attributes given as text descriptions of product, uses attributes based on expert description such as fashion, psychological age and attractiveness. The dataset has features of text, linear and nominal scales. The computational experiment shows the effectiveness of the proposed algorithm. The importance of features of the collection of data is defined. A software product demonstrating the recommendation system in action is developed.

Video on demand (VOD) service is widely used nowadays to watch online videos. Video on demand has evolved as a major Problem implementation for network. This research aims to design a suitable grid computing system to enhance performance of video on demand. The goal of using Grid computing system is to create the simple but large and powerful virtual computer with large collection of heterogeneous systems sharing resources. User can submit request for video to grid server without necessary knowing where it will be execute. It is the responsibility of the grid server to distribute request among several servers (clusters). The results we get optimize resource usage, load balance, and time delay for delivering video file. The MATLAB simulation program is designed to compare old server performance with Grid server performance focusing on time delay and system throughput, which decreases Video delay 60 % - 70%.

Fingerprint recognition is one of the most prominent biometric identification techniques. The features of fingerprint are broadly classified into three categories namely level 1, level 2 and level 3 features. In order to increase the recognition rate here the hierarchical matching technique is used which allows us to use all levels of features hierarchically. The level 3 features can be observed only in 1000 dpi images. So the matching is performed on a 1000 dpi fingerprint database captured using Hamster IV fingerprint scanner. At the first level of hierarchical matching the agreement between two orientation images is tested using dot product of the images. If an image is accepted at the first level it is brought to the second level of matching where minutiae details are tested. If an image crosses this stage as accepted then the decision will be taken as genuine otherwise the image will be given to the third level. The third level extracts the level 3 features such as pores and ridge contour points using Dynamic Anisotropic Pore Model (DAPM) and Mexican hat wavelet transform respectively. For each matched minutiae the pores and ridge contour points within the associated region of that minutiae will be matched. Based on the third level of matching a person is accepted as genuine or rejected as imposter.

Lifestyle in the modern society along with human behavior and thinking is changing dramatically with the advancement of technology, and the concept of a simple home is changing into a smart home. Home Security has been a major issue where crime is increasing and everybody wants to take proper measure to prevent intrusion, besides homes the system is also implemented in business premises and office. This project is residential intrusion detection and it detects the face by using image processing

The download time of website depends on various web components such as multimedia size, document size, program size and so on. The main objective of this paper is to Analyze Download time of University Websites in India and evaluating the quality of Website Download time Performance based on Download time Performance metric. The Download time of websites is measured in various grades viz., A, B, C, D, E, F etc., A 10 point metric for Download time performance is investigated based on Download time performance grades.

Separation of ions is a major problem in chemical industries. There are several techniques to separate ions, one of the technique is the use of electrokinetic valve. Electrokinetic valve is used to separate ions by applying external electric field. Earlier it is designed using CAD software. In this paper we present an approach of designing electrokinetic valve using COMSOL multiphysics in a miniature way.

In the modern cryptography, the problem of the generation of large primes is considered as an important issue to implement public key crypto-scheme such as RSA and so on. Since a great amount of computational resources is generally required to generate a large prime number, it takes considerable amount of resource especially in embedded systems. In general, the prime number generation is started from the random number generation. If the generated random number is passed a specified probabilistic primality test, the random number is tentatively considered as a prime number and applied to a public key crypto-system. In many primality tests, a number of modular exponentiations are usually required. Hence these algorithms are considerably slow. Here we are introducing a new prime generation technique using GA which can be used to generate large primes in more efficient way.

Bio-metric identification systems are widely used for unique identification of humans mainly is verification and identification .It is used as a form of identity access management and access control. Fingerprint identification is one of the best biometric technique. The biometric fingerprints identification are secure to use and mainly unique for every person and do not change in one’s lifetime. This paper is to identify each and every student of the department as well as the college.

Nowadays security has always been an important issue in the business world to ensure the integrity of content and transactions, to maintain confidentiality and privacy, and to make sure information is disseminated appropriately. But while dealing with contents encoded in XML, the hierarchical tree structured data have different confidentiality and integrity for different portions of the same content. Thus need dissemination approach specifically tailored to XML that addresses the issues of efficiency and scalability must however be provided by assuring at the same time security of contents and privacy of the parties acquiring and disseminating contents. The main objective is to provide solution to above problem using structural properties of tree based data model (such as XML document) and Document Object Model (DOM). This approach is based on notion of Encrypted Post Order Numbers (EPON) which is based on tree traversals and is defined using a random notation of such traversal number. It facilitates efficient dissemination of selected portion of content. By using structure based routing (SBR) scheme that prevents information leakage and assures that delivered content to user according to access control policies. The structure based routing framework facilitates the dissemination of contents with varying degrees of confidentiality and integrity in a network.