CLASSIFICATION OF CT IMAGE LUNG CANCER DISEASE USING HYBRID CLUSTERING AND DEEP LEARNING TECHNIQUES
European Journal of Molecular & Clinical Medicine,
2020, Volume 7, Issue 3, Pages 5475-5491
AbstractAmongst various cancers disease of human, lung cancer is considered to be the primary cause of cancer demise with greatest fall rapidity. Lung cancer is the unrestrained enlargement of irregular cells that begins off in one or both 2D CT images. Lung cancer could be detected by Computer Aided Diagnosis (CAD) test like Computer Tomography (CT) scanning as it gives more decoded image. The CT imaging is always preferred due to low radiation while compare to Magnetic Resonance Imaging (MRI). To classify various stages of lung cancer, digital image processing phases are used. The various phases of digital image processing techniques are used to classify different stages of lung cancer. Lung cancer detection and segmentation has various phases such as image preprocessing, segmentation, feature extraction and classification. The preprocessing technique is used to remove various noises and improve quality of image. Image preprocessing technique is carried out using 2D Adaptive Gabor Diffusion Filter (2D-AGDF) algorithm and Edge Preserved Contrast Limited Adaptive Histogram Equalization (EP-CLAHE) algorithm. Image segmentation technique is used to extract cancer pixels from CT image using Adaptive Mean Shift Threshold (AMST) methodology. The Gray Level Co-Occurrence Matrix (GLCM) feature extraction technique is used to calculate various features from the segmented image. The Deep Convolution Neural Network (DCNN) is used to classify the CT image whether it is normal or abnormal. The various experimental results and graphical representations prove that proposed methodology gives higher range of efficiency and accuracy in classifying lung cancer images.
- Article View: 49
- PDF Download: 96