Feature extraction and I-NB classification of CT images for early lung cancer detection

Abstract

The tremendous increase in health condition has fixed a novel confront to clinical routine of patients’ record about treatment, diagnosis and continuous follow-up with the assistance of data and image processing. It is probable to automate or support the radiologist for the diagnosis of cancer. Identification of lung cancer is performed by MRI, CT scanned image. CT scan shows significance totally and acts as a vast function in earlier diagnosis of cancer. So as to overcome certain shortcoming in selecting the feature and enhance classification, this investigation makes use of a novel evolutionary algorithm termed Accelerated Wrapper based Binary Artificial Bee Colony (AWB-ABC) Algorithm for efficient feature selection and Improved Naive bayes classifier for classifying the stages of cancer. Here, LUNA 16 dataset is considered as an input dataset, followed by the pre-processing of data for noise removal and image enhancement using MMSE. The morphological features are extracted from the pre-processed image, then the feature or nodule associated with the lung which has the great influence to cause cancer is to be selected, for this purpose, here AWB-ABC algorithm is utilized. At last, an Improved Naive Bayes (I-NBC) algorithm is used for efficient classification. The outcome shows a better trade-off between the prevailing and the proposed techniques. The accuracy of classification is improved by 98% using the proposed method. The simulation was carried out in MATLAB environment.