ICTACT Journal on Soft Computing

C. Bhuvaneswari ¹, P. Aruna ¹, D. Loganathan ²

Affiliations
1 Department of Computer Science and Engineering, Annamalai University, IN
2 Department of Computer Science and Engineering, Pondicherry Engineering College, IN

Abstract

Lung diseases are one of the most common diseases that affect the human community worldwide. When the diseases are not diagnosed they may lead to serious problems and may even lead to transience. As an outcome to assist the medical community this study helps in detecting some of the lung diseases specifically bronchitis, pneumonia and normal lung images. In this paper, to detect the lung diseases feature extraction is done by the proposed shape based methods, feature selection through the genetics algorithm and the images are classified by the classifier such as MLP-NN, KNN, Bayes Net classifiers and their performances are listed and compared. The shape features are extracted and selected from the input CT images using the image processing techniques and fed to the classifier for categorization. A total of 300 lung CT images were used, out of which 240 are used for training and 60 images were used for testing. Experimental results show that MLP-NN has an accuracy of 86.75 % KNN Classifier has an accuracy of 85.2 % and Bayes net has an accuracy of 83.4% of classification accuracy. The sensitivity, specificity, F-measures, PPV values for the various classifiers are also computed. This concludes that the MLP-NN outperforms all other classifiers.

Keywords

Feature Extraction, Multilayer Perceptron, Neural Networks, Bayes Net, Sensitivity, Specificity, F-Measure.

Full Text

     

R. Priya ¹, P. Aruna ¹

Affiliations
1 Department of Computer Science and Engineering, Annamalai University, IN

Abstract

Diabetic retinopathy (DR) is an eye disease caused by the complication of diabetes and we should detect it early for effective treatment. As diabetes progresses, the vision of a patient may start to deteriorate and lead to diabetic retinopathy. As a result, two groups were identified, namely non-proliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). In this paper, to diagnose diabetic retinopathy, three models like Probabilistic Neural network (PNN), Bayesian Classification and Support vector machine (SVM) are described and their performances are compared. The amount of the disease spread in the retina can be identified by extracting the features of the retina. The features like blood vessels, haemmoraghes of NPDR image and exudates of PDR image are extracted from the raw images using the image processing techniques and fed to the classifier for classification. A total of 350 fundus images were used, out of which 100 were used for training and 250 images were used for testing. Experimental results show that PNN has an accuracy of 89.6 % Bayes Classifier has an accuracy of 94.4% and SVM has an accuracy of 97.6%. This infers that the SVM model outperforms all other models. Also our system is also run on 130 images available from "DIARETDB0: Evaluation Database and Methodology for Diabetic Retinopathy" and the results show that PNN has an accuracy of 87.69% Bayes Classifier has an accuracy of 90.76% and SVM has an accuracy of 95.38%.

Keywords

Probabilistic Neural Network, Bayesian Classification, Support Vector Machine, Sensitivity, Specificity, Accuracy.

Full Text

     

M. Pauline ¹, P. Aruna ¹, B. Shadaksharappa ²

Affiliations
1 Department of Computer Science and Engineering, Annamalai University, IN
2 Department of Computer Science and Engineering, Shirdi Sai Engineering College, IN

Abstract

The Authors have proposed a model that first captures the fundamentals of software metrics in the phase 1 consisting of three primitive primary software engineering metrics; they are person-months (PM), function-points (FP), and lines of code (LOC). The phase 2 consists of the proposed function point which is obtained by grouping the adjustment factors to simplify the process of adjustment and to ensure more consistency in the adjustments. In the proposed method fuzzy logic is used for quantifying the quality of requirements and is added as one of the adjustment factor, thus a fuzzy based approach for the Enhanced General System Characteristics to Estimate Effort of the Software Projects using productivity has been obtained. The phase 3 takes the calculated function point from our work and is given as input to the static single variable model (i.e. to the Intermediate COCOMO and COCOMO II) for cost estimation. The Authors have tailored the cost factors in intermediate COCOMO and both; cost and scale factors are tailored in COCOMO II to suite to the individual development environment, which is very important for the accuracy of the cost estimates. The software performance indicators are project duration, schedule predictability, requirements completion ratio and post-release defect density, are also measured for the software projects in my work. A comparative study for effort, performance measurement and cost estimation of the software project is done between the existing model and the authors proposed work. Thus our work analyzes the interactional process through which the estimation tasks were collectively accomplished.

Keywords

Effort Adjustment Factor (EAF), Kilo Lines of Code (KLOC), Total Effort Multiplier (TEM), Scale Factors, Cost Drivers, Value Added Factors.

Full Text