Enhanced Lung Nodule Classification Using a Hybrid Clustering and Bayesian Approach

Lung cancer is a fatal disease that causes death in human beings at an early stage. For identification of disease, surgery planning medical image analysis finds a major role. Lung cancer is one of the dangerous diseases which results in the death of most of the human being because of uncontrolled cell growth. Automated lung cancer detection is very popular healthcare innovation technique as the conventional methods available for cancer detection are very painful. Though there are several methods that exist for segmentation with existing drawbacks so divergence exist in its prediction. A method is proposed with two sources for cluster construction: center homogeneity using Modified Expectation Maximization (MEM) Clustering algorithms and border homogeneity using Density-based clustering non-parametric algorithm (DBSCAN) algorithm and classification is based on Bayesian Classifier for accuracy and false positive rate enhancement.