Volume 4 Number 4 March 2015


Diagnosing Heart and Glaucoma Diseases using Retinal Vessel and OD Segmentation
P. Lekha, K Sudha

Abstract: Retina is responsible for capturing the visual and it triggers the nerve impulses in the brain. Retina is related to heart through the blood vessels which are connected to the arteries and veins in the heart. Blood vessels in the retina reflect the changes in the blood vessels of other parts of body like heart, brain, kidney etc., The six largest arteries and veins are measured using CRAE and CRVE which have strong correlation with stroke and heart diseases. Thus wrong identification of vessels leads to wrong diagnosis. Hence a post-processing step is introduced to vascular segmentation for identifying the true vessels. It models the segmented vascular structure as a vessel segment graph and the problem is formulated as finding the optimal forest. In addition to finding the cardio-vascular disease, in the proposed work, glaucoma disease is also identified. It identifies the various eye related infection by just segmenting the optical disk and cup using medial axis detection and vessel bends detection. It implements the cup boundary algorithm to find the cup to disk ratio. Based on this ratio the type of eye disease is detected.


Simultaneous Scheduling of Machines and AGVs in Flexible Manufacturing Environment with Maximization of Robust Factor Criterion
M. Nageswararao , K. Narayanarao, G. Rangajanardhana

Abstract: This paper examines the simultaneous scheduling of machines and two identical automated guided vehicles (AGVs) in a flexible manufacturing system (FMS).Optimum AGVs operation plays a crucial role in improving the performance of FMS. A genetic particle swarm vehicle heuristic algorithm (GPSVHA) is proposed and developed the code in JAVA to provide optimum sequence with maximizing the robust factor and minimization of mean tardiness and AGVs schedule for ten job sets and four lay-outs. The code will enhance the productivity, minimize the delivery cost and optimally utilize the entire fleet. the code provides better performance when compared with other algorithms, namely viz; sliding time window (STW), ulusoy genetic algorithm (UGA), abdelmaguid genetic algorithm (AGA), rao and reddy genetic algorithm (PGA), deroussi hybrid algorithm (DHA), chowdary genetic algorithm (CGA), hybrid genetic vehicle heuristic algorithm(HGVHA).


Performance Analysis of Java NativeThread and NativePthread on Win32 Platform
Bala Dhandayuthapani Veerasamy, G. M. Nasira

Abstract: The most important motivation for using a parallel system is the reduction of the execution time of computation-thorough application programs. To facilitate the development and analysis of parallel programs the performance measures are often used. Performance measures can be based not only on theoretical cost models but also on measured execution times for a specific parallel system. In this article, we selected different computer architectures and iterations to compare the performance results of NativeThread and Hybrid NativeThread with Thread class and Concurrent API as well to compare the performance results of NativePthread and Hybrid NativePthread with Thread class and Concurrent API. The overall performance improvements are five times faster than Thread class and Concurrent API.


Congestion Avoidance based on RC-MAC Protocol in Wireless Sensor Networks
G Velu, M B Suryatheja, C Ram kumar

Abstract: Application for wireless sensor network is notably different in its characteristics and requirement from a standard WLAN. When a critical event triggers a surge of data generated by the sensor, congestion may occur as data packet converge toward a sink, which causes energy waste, throughput reduction, information loss, hidden terminal problem and link failure in contention based MAC protocol. Due to hidden terminal problem the RC-MAC and CSMA/CA protocol sensing is used to reduce the channel contention and radio collision. The formation of a new tree for avoiding link failure-a MAODV routing protocol was introduced in this paper, and the performance of this protocol is measured using the above parameters. The demonstration of a near optimal throughput at each sensor and to achieve congestion avoidance in presence of a multicast routing towards a multiple sink is also done.


A Review on Approaches at Different Stages of Mammogram Processing
K Thangavel, R Roselin, R Subash Chandra Boss

Abstract: The aim of this paper is to review existing approaches of processing mammograms at different stages to detect breast cancer at the earliest. Moreover this paper helps to understand the different stages in mammograms and the already existing techniques in that area for further exploration. The review has been done in different stages namely mammogram preprocessing, segmentation, feature extraction, feature selection and classification in the recent years. The results obtained using different techniques are also reported.


Cardiac Biometrics: Human Identity Verification using PCG signals by Binary Decision Tree based SVM
K Lakshmi Devi, M Arthanariee

Abstract: Automatic Identity verification through Cardiac Biometrics is a profound area of Research where Heart sounds are analyzed, aiming in enhancing accuracy, thus reducing falsification. This paper examines the applicability of Biometric features of Heart sounds by analyzing the Phonocardiogram signals. Binary decision tree based Support Vector Machine Method is a new approach in the research. DWT results with different frequency bands that are smoothed by Multi-pass Moving Average Filters. Peaks are detected by Averaging Neighbors. Spectral features are extracted and clustered by HSOM. Rough Sets Theory (RST) selects the best features for classification. Binary Decision tree based SVM is used as a classifier for recognition and Identification.


Automatic Segmentation of Fetal Brain from MRI of Human
K Somasundaram, T Kalaiselvi, S P Gayathri, R Rajeswaran

Abstract: Fetal MRI is an essential tool for analyzing morphological changes of fetal brain structure. The automated methods developed for adult brain extraction are unsuitable for fetal brain extraction because of the differences in tissue types and tissue properties between adult and fetal brain. However, only few automated fetal brain segmentation methods are available. In this paper we propose a fully automatic method to extract fetal brain. The proposed method finds an ROI that encloses the fetal brain, using the anatomical geometry. An intensity threshold is computed using Otsu’s method, from which a binary image is obtained for the ROI. Using anatomical knowledge the fetal brain is extracted. Experiments were performed on clinical in utero fetal MR volume and the results are validated against manual segmentation and quantified in terms of Dice (D) similarity coefficient, Sensitivity (S), Specificity (Sp) and Hausdorff distance (HD). The results emphasize the robustness of the method.


An Efficient Product Hybrid Feature Classification on Opinion Mining using Ant Optimization Rule
P Saravana Kumar, A Vijaya Kathiravan

Abstract: Opinion mining refers to opinion retrieval with a set of searched results, produced for the customer’s depending on the product request demand. In general, opinion refers to a person’s perspective and mining is implied for identifying the result of the search query. Despite its promise, failure of clear product classification results in the ambiguity for obtaining additional product feature association. Also, several hybrid features related to the product were not classified with high efficiency rate. The list of product attributes classification (i.e., based on quality of products and specific feature selection of the products) with opinion mining algorithms was not investigated widely. To classify the product features based on the user behavior, an Ant Opinion Miner Classification Rule Learning (AOM-CRL) process is proposed in this paper. Initially, the process of Ant Opinion Miner helps to fetch the combination (i.e., hybrid) product features based on the user behavior. The Ant Opinion Miner involves the pheromone initialization, updation and fitness function for efficient classification of hybrid features. Pheromone initialization starts the process based on the user behavior with opinion patterns based classification on product features. Then the pheromone updating process is carried out with the same behavior for a set of hybrid features of the products. Both the initialization and updating uses the user behavior information to easily adopt the classification rule. The fitness function is employed in AOM-CRL process to produce quality result on classification process. In addition, different set of hybrid feature classification also produces high quality function value through ant-miner based classification rule. This ant-miner based classification rule integrates the Enhanced Rapid Repeated Attribute Reduction algorithm to improve the classification accuracy on opinion mining with minimal processing time. Experiment is conducted with classification rate on user opinions, false positive rate and precision rate on product result.


A Pragmatic Study of LEACH and its Descendant Routing Protocols in WSN
G Devika, AshaGowda Karegowda

Abstract: A wireless sensor network (WSN) is a wireless network consisting of spatially distributed autonomous sensors to monitor physical or environmental conditions. The battery power in these sensor nodes plays an important role in increasing the lifespan of the nodes. Hierarchical routing protocols are the best known protocols to minimize the energy consumption. Low-Energy Adaptive Clustering Hierarchy (LEACH) is a classical cluster based routing protocol for WSNs having good performance, energy efficiency, and effective in prolonging the network life time by consuming a small percentage of the total dissipated energy in the system. This paper surveys the state-of art of different hierarchical routing protocols that have been developed from the LEACH. This paper highlights some of the drawbacks and issues of LEACH; how these issues are conquered by the descendants of LEACH. Assessment of descendents of LEACH routing protocol in briefed in terms of scalability, self-organization, and distribution of nodes, network control, hop count, energy efficiency, and use of location information.


Analysis of Biclustering Algorithm using Synthetic Data
S Guru, T Marimuthu, R Lawrance

Abstract: Biclustering technique is an alternate approach for standard clustering methods, which helps to identify the local structures from the gene expression data. These local structures provide information about main biological functions that linked with the physiological states. Biclustering method clusters the rows and column concurrently. Most of the biclustering algorithm works based on the various scores like mean square residue, variance, co-variance etc. For calculating these scores many of the algorithms follow Cheng and Church algorithm. In this work, we review the Cheng and Church algorithm and demonstrate the working procedure of the same. We have formed the synthetic data for showing the results of Cheng and Church algorithm. The results of this work clearly derived the constant and additive bicluster patterns.