ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

COMPUTER AIDED DIAGNOSIS OF LUNG DISEASES USING KNN CLASSIFIER

Journal: International Journal of Engineering Sciences & Research Technology (IJESRT) (Vol.7, No. 6)

Publication Date:

Authors : ;

Page : 199-208

Keywords : K-Nearest Neighbor; Support Vector Machine; Random Forest.;

Source : Downloadexternal Find it from : Google Scholarexternal

Abstract

Lung diseases like pneumothorax and pleural effusion are detected and diagnosed using KNN classifier. This paper proposes a real-time detection algorithm to classify the lung diseases. Initially, a real-time image is collected from the public database. Since the collected images are prone to noise this is removed by the preprocessing techniques. The preprocessed image is further segmented by thresholding method. The inner white region and outer white region of the CT image is removed using image clear border. Then the centroid and the heights of each lung are found. The segmented image is overlapped with the original gray image. The texture features are extracted from the segmented images. The extracted features are energy, contrast, correlation, homogeneity and area which help to classify the CT images. KNN classifier is used to classify the CT image either as normal lung or lung affected with Pneumothorax or Pleural Effusion. Experimental analysis is performed using normal lung images and diseased lung images using KNN, RF (Random Forest) & SVM (Support Vector Machine) classifier. The accuracy and precision of KNN classifier is 90% & 100%, RF classifier is 81% & 80% and SVM classifier is 72% & 75%.

Last modified: 2018-06-08 19:39:59