A Computer Aided Diagnosis for detection and classification of lung nodules

Lung cancer has been the deadliest among all other types of cancer. Early detection of cancer is required to increase the survival rate of cancer patients. Our purpose is to develop an efficient Computer Aided Diagnosis (CAD) for detection of lung nodules from parenchyma region of lung and classify the nodule into either cancerous (Malignant) or non-cancerous (Benign). The proposed system consists of following steps: i) the image taken is enhanced initially and then the region of interest is cropped, where the user can select the area to be cropped. ii) Morphological operation is performed to suppress the blood vessels and enhance the nodules. iii) Nodules are identified by labeling. iv)Those identified nodule's features are extracted. v) Neural networks are implemented as the classifiers that works basically, based on the features extracted. The proposed work was able to detect the lung nodule that falls in close proximity to the lung wall. The system is able to achieve an overall accuracy of 92.2%.