New user registration
Event organizer registration
Login via your institution account
Login via your organization account
Login via your publisher account
Login using your new user credentials
Forgot username or password?
Subscribe to SciArchives
Indian journal of innovations and developments
Publisher:Indian Society for Education and Environment
Editor in chief:Prof.Natarajan Gajendran
Your selection(s) could not be saved due to an internal error. Please try again.
Browse by Volumes
Advanced Medical Image Compression With 2-D Maximum Entropy Method And Hybrid Compression Concepts
Author:A.J. Rajeswari Joe, N. Rama.
Volume: 5 | Issue no: 8-2016 | Pagination: 1-7
Objective: In medical image processing, storage and transmission of image pixels is a vital problem since diagnosis is a continuous process. Methods:The 2D maximum entropy method segmentation followed by compression provides good results in the field of Medical image processing which has storage and transmission problems.The proposed algorithm is created using 2-D maximum entropy method segmentation and followed by Hybrid image compression.This proposal introduces an advanced compression method which combines both lossy and lossless image compression techniques. The clinical part of the image is segmented with 2-D maximum entropy thresholding and lossless techniques are applied to compress this part. Lossy image compression technology is implemented in other part of the image that is with the background picture. Run Length Encoding is applied to the resultant data to produce the compressed image. Findings:The Advantage of using 2-D Maximum entropy method is that it considers both the grey information and neighbouring information. It can be able to produce good results even the image's signal to noise ratio (SNR) is low. Among lossy and lossless image compression methods lossless method is preferred in medical image processing since it has to protect important clinical information. Lossless compression can be able to produce lower compression ratio than lossy because it reduces the size of image only to certain limit. The experimental result shows that the proposed method provides good compression ratio and PSNR measures. Novelty/Improvements:Future enhancement of this algorithm is that the neural network algorithm such as selforganizing feature map can be used to find out the threshold value automatically.
Som Based Clustering For Detecting Bacterial Spot Disease In Tomato Field
Author:X.Agnes Kala Rani, R.Nagaraj.
Volume: 5 | Issue no: 7-2016 | Pagination: 1-8
Objectives: The main objective of introducing SOM based clustering method is to improve the classification accuracy and detection of bacterial spot disease in tomato field. Methods: There are various image processing methods used to identify disease and severity of disease in plants. One of such methods uses visible spectrum Images for automatically detecting and classifying the severity of bacterial spot in tomato fields. Centroid-based K-means clustering was widely used for automatic segmentation. Findings: Plant diseases are one of the major responsibilities for economic degradation in the agricultural industry. So regular monitoring of plant health and early detection of disease causing pathogens are required for minimizing disease spread and assist effective management practices. Centroid-based K-means clustering for segmentation always does not chose centroids that provide best results and also different initial set of centroids affect the shape and effectiveness of the final cluster. Application/Improvements: To overcome the limitations of Centroid-based K-means clustering, Self-Organizing Maps (SOM) is introduced for achieving effective classification result and to improve the detection performance.
Hydrodynamic Studies In Miniature Rotating Disc Contactor
Author:Mahesh Agrawal, S. Balasubramonian, Shekhar Kumar.
Volume: 5 | Issue no: 7-2016 | Pagination: 1-8
Objective: To employ the use of rotating disc contactor to extract leftover traces of Uranium and Plutonium from used fuel cells. Methods/Statistical analysis: The hydrodynamic variables were studied under no mass transfer condition between 0.01N Nitric Acid (aqueous phase) and a solution of 30% Tri-Butyl Phosphate and 70% Dodecane (organic phase) in a miniature rotating disc contactor. There were two rounds, first with aqueous as continuous phase (organic as dispersed phase) and aqueous as dispersed phase (organic as continuous phase).The drop size analysis was carried out using Image J software and Microtrac S3500. Findings:The empirical behavior of different hydrodynamic variables such as hold up, characteristic velocity, flooding and drop size, were observed by varying the rotor speed and inlet flow rate of the continuous phase. Consistency of empirical results were checked with standard literature correlations. Finally, discrepancy in the theoretical and experimental values were identified and explained. Improvements: This project leaves an opportunity for development of a unified correlation to obtain the value of hold up in no mass transfer condition for different rotor speeds.
A Survey On Various Data Dissemination Approaches In Vehicular Ad Hoc Network
Author:Dr.R. Priya, Sulfath P.M.
Volume: 5 | Issue no: 7-2016 | Pagination: 1-4
Objectives: The main objective of this work is to survey various techniques used in the field of data dissemination in Vehicular Ad hoc NETwork (VANET). Methods: There are different approaches used to improve data dissemination in vehicular ad-hoc network. Such approaches are context-aware information dissemination, Space-time network coding, geo-based Named data networking forwarding strategy, Route-based data pre-fetch model and etc. Findings: VANET has become more popular in the areas of vehicular transportation systems. Data dissemination which is the base of communication plays a major role in VANET system. Various data dissemination technologies were developed for avoiding the issues such as data routing, vehicle mobility and network security. This paper provides brief explanation about several data dissemination techniques and compares their performance based on the output parameters, merits and demerits. Comparison result shows that, the prefetching based data dissemination technique outperforms than other techniques. Application/ Improvements: the result of this work shows that prefetching based data dissemination technique provides better result than the other dissemination techniques.
A Survey On Pseudonym Management Scheme In Vehicular Ad-Hoc Networks
Author:T. Sivakumar, Nidhin A.S.
Volume: 5 | Issue no: 6-2016 | Pagination: 1-4
Objectives: To evaluate various pseudonym management schemes in vehicular Ad-hoc Network (VANET) for enhancing confidentiality or ambiguity. Methods: Vehicular Ad-hoc Network (VANET) is one of the categories in mobile ad-hoc networks (MANET), which is created for the intelligent transport system (ITS).VANET is utilizes to permit vehicles in terms of self-organized network system not including the requirement of the enduring transportation. The pseudonym management is the technique amongst the road side units (RSU) and vehicles for providing the requirements of the vehicles. The major issue in VANET is anonymity which is achieved through pseudonym management technique. There are different approaches developed to support the anonymity in VANET. Findings: This paper investigates detailed information of different approaches and finally compared their effectiveness. Application/Improvements: The finding of this work shows that Pseudonym shuffling process based scheme is better than other techniques.
A Survey On Different Approaches For Sentiment Analysis Of People
Author:Sivakumar T, Amitha Joseph.
Volume: 5 | Issue no: 6-2016 | Pagination: 1-4
Objective: To analysis different techniques and approaches for sentiment analysis to know the user opinion about a product or event or service that helps to improve an organization. Findings: In an emerging network every company wants to know users opinion about their product or service. Each and every user has different views about the product or service and their views are expressed through reviews. The analysis of such opinion from different users plays an important role in the growth of a company. The opinions are expressed through reviews in the natural language. Sentiment analysis is a process used to identify emotions, opinions and evaluations and it also predict the orientation of sentiment whether the sentiment is positive, neutral or negative opinion based on the words or sentences expressed in the reviews. Sentiment analysis is otherwise called as opinion mining. In this paper various techniques and approaches for sentiment analysis are analysed and finally compared their effectiveness through parameters like accuracy, precision, recall and F-measure values. Results: In this paper various techniques for sentiment analysis techniques are compared through parameters to prove unsupervised approach at aspect level for sentiment analysis is better than other techniques. Application/Improvements: The finding of this work shows that unsupervised approach at aspect level for sentiment analysis is better than other techniques.
A Survey On Various Approaches For Taxonomy Construction
Author:S. Sritha, B. Mathumathi.
Volume: 5 | Issue no: 6-2016 | Pagination: 1-6
Objectives: To analysis different approaches for taxonomy construction to improve the knowledge classification, information retrieval and other data mining process. Findings: Taxonomies learning keep getting more important process for knowledge sharing about a domain. It is also used for application development such as knowledge searching, information retrieval. The taxonomy can be build manually but it is a complex process when the data are so large and it also produce some errors while taxonomy construction. There is various automatic taxonomy construction techniques are used to learn taxonomy based on keyword phrases, text corpus and from domain specific concepts etc. So it is required to build taxonomy with less human effort and with less error rate. This paper provides detailed information about those techniques. Methods: The methods such as lexico-syntatic pattern, semi supervised methods, graph based methods, ontoplus, TaxoLearn, Bayesian approach, two-step method, ontolearn and Automatic Taxonomy Construction from Text are analyzed in this paper. Application/Improvements: The findings of this work prove that the TaxoFinder approach provides better result than other approaches.
Improved And Ensemble Methods For Time Series Classification With Cote
Author:Yamunadevi S, Dr.K.Sasi Kala Rani, Pavithra M, Priyanga M
Volume: 5 | Issue no: 6-2016 | Pagination: 1-9
Background/Objectives: To classify the time series data efficiently by introducing Collective of Transformation-Based Ensembles method (COTE). Methods/Statistical analysis: In existing scenario, the method is introduced named as COTE. It is mainly used for increasing the classification accuracy than preceding research. Another algorithm is named as Time series classification (TSC) which is used for transformation process which is based on comparative features. COTE contains classifiers constructed in the time, frequency, change, and shapelet transformation domains combined in alternative ensemble structures. However it has issue with transformation process and hence accuracy of the classification is reduced significantly. To avoid this issue introduced the concept called as run length transformation to improve the classification accuracy higher than existing system. Findings: The run length algorithm is improved along with genetic approach to produce the optimal features. In this scenario, the measures are considered as similarity coefficient, likelihood ratio and dynamic time warping (DTW). Based on the modified k- nearest neighbor distance concept the speed is increased and classification accuracy is improved prominently. Improvements/Applications: From the experimental result we can conclude that our proposed scenario yields better classification performance rather than existing scenario.
A Survey On Prediction Of Brain Hemorrhage Using Various Techniques
Volume: 5 | Issue no: 6-2016 | Pagination: 1-3
Objectives: The main objective of this work is to predict Subarachnoid haemorrhage (SAH) using machine learning techniques and analyzing the classification performance of various existing machine learning algorithms. Methods: Diagnosing theSubarachnoid haemorrhage can be done efficiently by various machine learning techniques. Purpose of using Machine learning technique is to focus on factors that influence the prediction performance. Findings: Subarachnoid haemorrhage is a stroke which is recognised by the occurrence of blood in subarachnoid space. Diagnosis of such potential disease becomes more important in the medical research area. Most widely used data mining methods for prediction tasks are decision rules, na´ve Bayesian classifiers, support vector machines, Bayesian networks, and nearest neighbors. Some of the methods namely boosting, bagging and genetic algorithms have limited usage in the prediction. Application/Improvements: The finding of this work shows that random forest classifier provides effective classification result than other machine learning techniques.
A Survey On Different Similarity Join To Improve Clustering, Classification And Similarity Search
Author:C.P. Rushida, V R Nagarajan.
Volume: 5 | Issue no: 5-2016 | Pagination: 1-6
Objectives: To analysis various similarity join techniques to improve the data mining process. Findings: Similarity join is an evaluation of similarity between any two objects. Many applications such as data cleaning, data integration, near duplicate detection and all data mining process can extensively benefit from the similarity join measure. Thus the similarity join can be performed between objects or strings or nodes etc. It finds all pairs of objects whose similarity is not smaller than the similarity threshold. There are different techniques and approaches are used to find the similarity join between objects in homogeneous information network. This paper provides detailed information about the different similarity join techniques. Results: In this paper various similarity join techniques are compared through parameters to prove path based similarity join is better than other techniques. Application/Improvements: The findings of this work prove that the path based similarity join provides better result than other approaches.
Contact and support
Terms and conditions
Frequently asked questions
Copyright © 2015
All rights reserved. SciArchives.