Nodule segmentation of lung CT image for medical applications

Abstract

Lung cancer increases number of deaths worldwide. This motivates the detection of Cancer in the early stages. In order to detect cancer in the various stages segmentation of initial lung portion and in turn, nodules is primary and very important step in diagnosing the cancer using Computer Vision (CV) and Machine Learning (ML) techniques. This paper contributes to the CV based cancer diagnosed system. This paper proposes cuckoo search optimization technique to optimize the initial segmentation of the lung portion, and then Active Contour (AC) technique has been applied to segment the nodules from the segmented lung image. In order to fine turn the post- processing after the nodule segmentation operation, Markov Random Field (MRF) technique has been applied. The results of the experimentation of the proposed method of lung and nodule segmentation have been reported in the results and discussion section.