Author Country (or Countries)



This paper proposes an intelligent approach for the development of a new support system to improve the performance of Computer Aided Diagnosis for automated pulmonary nodule identification on Computed Tomography images which is Digital Imaging and Communications in Medicine format. The first step in diagnosis of any abnormality in lung region, is to acquire a Computer Tomography image, a non-invasive procedure. The digital format of the image is highly portable, hence the extraction and sharing of valuable knowledge. The large number of related images pose a challenge in coherence and consequently arriving at conclusion. The CAD system has been designed and developed to segment the lung tumour region and extract the features which is of region of interest. The Detection process consists of two steps, namely Lung segmentation and Feature extraction. In segmentation of lung region K-means, Watershed and Histogram based algorithms is implemented. The extracted features and the label of the corresponding ROI were used to train a neural network . Finally , these properties are used to classify lung tumour as benign or malignant. The main objective of this method is to reduce false positive rate and to improve the access time and reduce inter-observer variability.

Digital Object Identifier (DOI)