الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 121 |  |  |
| | | from | 121 | | |
Abstract
A reliable computer-aided system for the automatic classification of different digitized computed tomographic (CT) images of the human’s lung abnormalities (Normal, Asthma and emphysema) is presented. The dataset was made available by collaboration between the ELCAP and VIA research groups. The proposed system consists of four main steps. First, the raw CT chest images were denoised to improving the image quality. Second, the human’s lungs were segmented from background and surrounding organs in the CT chest images. Third, five sets of features are extracted from each segmented lung image; statistical parameters, morphological features, texture features derived from gray level co-occurrence matrix (GLCM), Fourier descriptors and wavelet coefficients. Finally, the set of distinct features, checked by feature selection, are inputted to a set of artificial neural network (ANN) based classifiers. As a result of extensive comparative study; it has been found that the proposed system gives quite satisfactory detection performance. It has been found that the correct classification rate (CC) reaches 98.67% for the chosen database when ANN classifier using features derived from the discrete wavelet transform (DWT). Finally, data fusion techniques were used to improve the correct classification rates using two schemes of data fusion at the decision level. The classification correct rate reaches 99.947%. It has been proved that the results are satisfactory and higher than those reported by other researchers.