Entity

Time filter

Source Type


Kanchana A.,Mahendra Engineering College for Women | Arumugam S.,Nandha Engineering College Erode
International Journal of Applied Engineering Research | Year: 2015

Palmprint based biometric recognition has proven to be a reliable source of pattern matching between the tests and training images with variety of features. Recently, the Palm print biometric system is fast and reliable on correct matches but longer key based matching was not exhibited. The shorter key distribution is not effective on detecting the correlation pattern matching output. On the other hand, managing the non-linear deformation palm print is a crucial task since the matching rate is affected especially on minutiae, orientation, and density mapping. A better classifier index and matching method is needed to be employed on the palmprint biometric recognition system. In this paper, Diffie–Hellman-Merkle (DHM) Longer Key Exchange based Palmprint Pattern Matching method is employed. DHM based palmprint pattern matching comprises of three processes. Initial process, palmprint image classification is carried out using the Mahalanobis distance based classifier. Then the classified palmprint image is segmented based on the Higher Order Neighborhood Statistical (HONS) approach. Proposed HONS approach in the second process performs the morphological palmprint image segmentation on different non-linear deformation texture variance. Final processing step is the decision control process (i.e.,) pattern matching on minutiae, orientation, and density. Diffie–Hellman-Merkle longer key set is used to exchange between the two ends of the communication channel to improve the matching rate. The pattern matching rate in Diffie–Hellman-Merkle method essentially consists of finding the best matching between the template (i.e., stored palmprint image in database) and test user palmprint image. Experiment is conducted on factors such as false rejection rate, palmprint matching rate, rate of classification errors. © Research India Publications. Source


Stanly Jayaprakash J.,Mahendra Institute of Technology Mahendrapuri | Arumugam S.,Nandha Engineering College Erode
International Journal of Applied Engineering Research | Year: 2014

In recent years, hand based biometrics has attracted considerable attention. With a surge in the growth of e-commerce applications, the need for reliable system for user identification and enhanced security in hand based biometrics is acutely felt. Hand based biometric schemes provides several types of identification methods by the way of using finger print, plamprint, hand geometry and hand vein. Existing systems used the posture and position of hands using pegs for user identification. But, such approaches are difficult when group of users are chosen for experiments. Since the anatomy of the human hand is much complicated, the image pattern formation is highly unique. To address these existing deficiencies in the biometric system, it was proposed to develop an efficient finger print biometric authentication system, named Contour Identity Cross Detection technique (CICD). The evaluation of contour identity based finger print recognition is done using cross detection of the finger print contours across the sample finger print image. The CICD technique provide Identity Cross Detection algorithm for locating crossing regions by analyzing the angle of contour occurrence and by plotting the 4- connected lines between the distinct contour points for accurate matching. The CICD technique is able to the evolve an efficient biometric authentication system even in dynamic contours of the finger print sample. The CICD seeks to improve the matching performance, with use of global knowledge of finger prints. Simulation experiments were done for performing the finger print authentication system under various conditions to maintain the utmost security. © Research India Publications. Source


Arumugam S.,Nandha Engineering College Erode
European Journal of Scientific Research | Year: 2012

The fresh leaves of betelvine are consumed by about 20-25 million people in the country. It is cultivated following the traditional methods in India on about 75,000 hectares with an annual production worth about Rs 1000 million. There are some diseases infected in the entire plantation without any early symptoms of the diseases. This paper aims is to study the different types diseases infected in the betelvine plants and also procedure for to identify diseases early infected stage using digital image processing and pattern recognition techniques. The digital image of the betelvine leaves at different stages of the disease is collected from different plants using a high resolution digital camera and stored JPEG format. The image analysis of the leaves is done using the image processing toolbox in MATLAB gives the standard patterns of the digital images. The thresholding was done to remove the back ground. Using RGB encoding technique the red, green, blue components of the preprocessed image were separated which forms the pattern to be compared. These patterns and images of various healthy betelvine leaves at different stages in various days are collected and stored in the system. To compute the mean, median and standard deviation values for all healthy and infected leaves and calculated values are stored in the system. For the test leaf, compute mean, median and standard deviation values and compare all the stored values. To identify the diseased leaf affected by disease using digital image processing and pattern recognition techniques. © EuroJournals Publishing, Inc. 2012. Source

Discover hidden collaborations