Detection and classification of lung cancer MRI images by using enhanced k nearest neighbor algorithm

Abstract

Objectives: To detect and classify the malignant cancer tissues and benign cancer tissues in MR lung cancer images by using k nearest neighbor mining algorithm. Methods/Statistical Analysis: In this paper, the Enhanced K Nearest Neighbor (EKNN) algorithm is executed to identify the lung cancer images. The k nearest neighbor technique is an important method of data mining algorithms. Findings: This work implicates four stages such as pre-processing, feature extraction, ­classification and detection of cancer tissues. In preprocessing stage, morphological process is used to filter the irrelevant noisy data in images. In the second phase, statistical and discriminator algorithm is used to extract the images. In the last stage, the improved k Nearest Neighbor (EKNN) method has been used to classify and identify the cancerous tissues in MRI images. The detection of cancer tissues and classification is done by executing four steps of Enhanced k Nearest Neighbor method which are measuring the Euclidean distance value, determining the k value, calculating the minimum distance and detecting the cancerous cells. Improvements/Applications: The experimental study with enhanced k nearest neighbor method shows better and promising classification result for classifying benign and malignant tissues.  Keywords: Geometrical and Statistical Properties, Image Classification, Image Mining, MRI Images, k Nearest Neighbor, Morphological Method