PSN Engineering College

Tirunelveli, India

PSN Engineering College

Tirunelveli, India

Time filter

Source Type

Hariharan G.,PSN Engineering College | Rajaram M.,Anna University
Asian Journal of Information Technology | Year: 2016

Classification of medical imagery is a difficult and challenging process due to the intricacy of the images and lack of models of the anatomy that completely captures the possible deformations in each structure. Cervical cancer is one of the major causes of death among other types of the cancers in women world wide. Proper and timely diagnosis can prevent the life to some extent. Therefore we have proposed an automated reliable system for the diagnosis of the cervical cancer using texture features and machine learning algorithm in PAP smear images, it is very beneficial to prevent cancer also increases the reliability of the diagnosis. In this research study, we have developed, multi class cervical cancer severity analysis system based on hybrid texture features and hybrid RBF kernel based support vector machine using PAP smear images. Two major contribution of the proposed system is feature extraction and feature classification. In feature extraction, multiple features are extracted using texture features and Gabor filter based orientation image. This system classifies the PAP smear cells into anyone of four different types of classes using RBF-SVM. The performance of the proposed algorithm is tested and compared to other algorithms on public image database of Herlev University Hospital, Denmark with 452 PAP smear images. The overall classification accuracy of the proposed hybrid RBF-SVM is 96.8% but the existing methods RBF and SVM produce 91.32 and 94.32%, respectively. © Medwell Journals, 2016.


Chitradevi A.,PSN Engineering College
Current Medical Imaging Reviews | Year: 2016

Image segmentation is considered to be the most important practical aspect of image processing. It is bethought to have its application in medical imaging and also it acts as a clinical diagnostic tool. Medical image segmentation (MIS) is facilitated by automating the depiction of anatomical structures and other region of interest. In case of Computer Aided Diagnosis, MIS is considered to be an initial and essential step. Accuracy of image segmentation algorithms are focused more behind the success of any medical image analysis. Whatever may be the application either in radiotherapy planning or clinical diagnosis and treatment, exact segmentation of medical images are cared. Hence, it remains challenging, unsolved, and sometimes seems to be a complex task too. Various MIS algorithms have emerged, which are not suitable for all images. In this paper, different MIS approaches are categorized with their sub methods and sub fields. Recent techniques for every category are also discussed and the comparison of these approaches with pros and cons is summarized. Using these techniques to develop a new hybrid algorithm will be of very much use in medical diagnosis. © 2016 Bentham Science Publishers.


Vijaya Kumar B.,PSN Engineering College | Balakumar B.,M. S. University of Baroda | Raviraj P.,Coimbatore Institute of Technology
International Review on Computers and Software | Year: 2014

Uncontrollable Accurate brain tissue segmentation from magnetic resonance (MR) images is an essential step in quantitative brain image analysis. However, the existence of noise and intensity inhomogeneity in brain MR images, many image segmentation algorithms suffer from limited accuracy. In this paper, we have assume that the local image data within each voxel's neighborhood satisfy the Gaussian Mixture Model (GMM), and thus propose the Fuzzy Local GMM (FLGMM) algorithm for automated brain MR image segmentation. This algorithm estimates the segmentation result that maximizes the posterior likelihood by minimizing an objective energy function during which a truncated Gaussain kernel function is used to impose the spatial constraint and fuzzy memberships are employed to balance the contribution of each GMM. We compared our algorithm to progressive segmentation approaches in both synthetic and clinical data. An experimental result shows that the proposed algorithm will mostly overcome the difficulties raised by noise, low contrast, and bias field, and considerably improve the accuracy of brain MR image segmentation. © 2014 Praise Worthy Prize S.r.l.-All rights reserved.


Vijayaraghavan R.,Bharathiar University | Loganathan A.,Manonmaniam Sundaranar University | Rajalakshmi D.,PSN Engineering College
Journal of Testing and Evaluation | Year: 2013

A lot sensitive sampling plan is an inspection procedure that involves a single sample with a zero acceptance number, and it is useful in the areas of compliance and safety-related testing. It is regarded as a consumer-oriented plan and is applied for an individual isolated lot, giving protection in terms of the limiting quality level. A major disadvantage of the plan is that it creates the possibility of the rejection of good lots because of the severity of the acceptance criteria involved in the plan. In this paper, a modified lot sensitive procedure that uses the repetitive group sampling procedure with small acceptance numbers is proposed. It can be considered as an alternative to the lot sensitive single sampling plan. The selection of the plans providing protection to the producer and the consumer in terms of acceptable and limiting quality levels with associated risks is discussed through illustrations. Tables for determining the sample size and the points for the plot of the operating characteristic curves of the plans are constructed using f-binomial approximation to a hypergeometric distribution. Copyright © 2013 by ASTM International all rights reserved.


Kumar K.M.,PSN Engineering College
2014 IEEE International Conference on Computational Intelligence and Computing Research, IEEE ICCIC 2014 | Year: 2015

Analysis of structural changes in the brain through magnetic resonance imaging can provide useful data for diagnosis and clinical supervision of patients through dementia. While the degree of sophistication reached by the MRI equipment is high, the quantification of tissue structures and has not yet been completely solved. Segmentations that these teams now allow those structures fail where the edges are not clearly defined. In this paper a method of automatic segmentation of MRI brain images based on the use of Generalized Regression Neural Networks using genetic algorithms for parameter settings is presented. The network is trained from a single image and classifies the rest of them when the MRI images were acquired with the same protocol. A method of measuring the progressive atrophy and possible changes before a therapeutic effect should be essentially automatic and therefore independent of the radiologist. © 2014 IEEE.


Natchadalingam R.,PSN Engineering College | Somasundaram K.,Jaya Engineering College
ARPN Journal of Engineering and Applied Sciences | Year: 2015

Grid networking is an aggregation of geographically dispersed computing, storage, and network resources, coordinated to deliver improved performance, higher quality of service, better utilization, and easier access to data. It enables virtual, collaborative organizations, sharing applications and data in an open, heterogeneous environment. Scheduling is the process that selects which job in the queue should be considered next. Grid Scheduling is the process of making scheduling decisions involving allocating jobs to resources over multiple administrative domains. The goal of scheduling is to minimize the make-span by finding an optimal solution. In the present Grid Networking environment, the scheduling approaches. In a Grid Networking environment there are many more constraints that would make the job scheduling problem more complicated. The issues in the existing system are the dynamic environment of the Wireless Grid makes necessary the use of sophisticated mechanisms for resource discovery and selection. Task monitoring and check pointing is difficult in dynamic environments. In this paper we have proposed a Task Deferment Algorithm, using activates strategy system to effectively allocate the resources to the tasks by performing sort-out. If a task does not continue its execution due to disconnectivity of resources, the resources for that task will be provided immediately next from the task which has finished its execution. © 2006-2015 Asian Research Publishing Network (ARPN).


Natchadalingam R.,PSN Engineering College | Somasundaram K.,Jaya engineering college
International Journal of Applied Engineering Research | Year: 2014

Grid networking is an aggregation of geographically dispersed computing, storage, and network resources, coordinated to deliver improved performance, higher quality of service, better utilization, and easier access to data. It enables virtual, collaborative organizations, sharing applications and data in an open, heterogeneous environment. Scheduling is the process that selects which job in the queue should be considered next. Grid Scheduling is the process of making scheduling decisions involving allocating jobs to resources over multiple administrative domains. The goal of scheduling is to minimize the make-span by finding an optimal solution. In the present Grid Networking environment, the scheduling approaches for VM (Virtual Machine) resources only focus on the current state of the entire system. Most often they fail to consider the system variation and historical behavioral data which causes system load imbalance. In existing system is based only on future load prediction mechanism. Based on this factor VM resource allocation is done. During this VM migration, there is no suitable criterion for unique identification and location of VM that means which VM is migrated and where to be migrated. In this paper a Grid Booster Algorithm is used. In this system VM allocation is based on resource weight [a value indicates capacity of each resource]. Based on these weights a VM resource allocation mechanism has proposed, which is considering both resource weight and future prediction. © Research India Publications.


Vijayaraghavan R.,Bharathiar University | Loganathan A.,Manonmaniam Sundaranar University | Rajalakshmi D.,PSN Engineering College
Journal of Testing and Evaluation | Year: 2014

Acceptance sampling schemes are frequently used when finished products or materials are supplied by the producer for making decisions about individual lots or group of lots by examining sample information of such products or material. A sampling scheme is a set of plans that are usually specified by certain parameters such as sample size(s) and acceptance number(s). The tightened-normal-tightened (TNT) sampling scheme is a scheme that incorporates two single sampling plans called normal and tightened sampling plans, having the same sample size, but with different acceptance numbers along with the rules for switching between the plans. The determination of parameters of a sampling scheme for specified requirements on its operating characteristic curve providing protection to the producer and consumer is termed as designing of the scheme. This paper presents an iterative procedure for designing TNT schemes for two specified points on the operating characteristic curve, namely, acceptable quality level and limiting quality level associated with the producer's risk and the consumer's risk, respectively. Tables providing the parameters of the schemes for various values of acceptable quality level and limiting quality level under the conditions of binomial and Poisson distributions are developed and presented. The efficiency of such schemes over the schemes developed by the unity value approach is demonstrated with illustrations. Copyright © 2014 by ASTM International.

Loading PSN Engineering College collaborators
Loading PSN Engineering College collaborators