Lung Cancer Segmentation on CT Images Using Fast Fuzzy C Means Clustering and Adaptive Mean Shift Threshold Segmentation

K. Dayanandhan and Dr.S. Maflin Shaby

In the recent past, bio-medical imaging is playing a significant role in diagnosis of various types of diseases occurred in human body. Based on various types of cancers, lung cancer is the most common cancer keeping the maximum mortality rate. The Computer Tomography (CT) technology is used to scan the human organs for visually understanding issues occurred on it. The CT scanning is applied for detection of lung cancer as it gives specific meaning of tumor in the human body and tracks its enlargement. In this research, a novel methodology is proposed segmentation and classification for lung cancer using CT images. The CT image is preferred due to its low radiation to patients while compared to Magnetic Resonance Imaging (MRI). The CT image is basically poor and highly distorted due to various types of noises. The CT is applied to preprocessing technique in order to remove various types of noises and to improve the quality in terms of contrast and brightness. The 2D Adaptive Gabor Diffusion Filter (2D-AGDF) algorithm is proposed to remove various types of noises. The contrast and brightness are improved by applying Edge Preserved Contrast Limited Adaptive Histogram Equalization (EP-CLAHE) algorithm in order to preserve edges of CT images. The Fast Fuzzy C Means technique is a hybrid clustering approach that is applied for clustering of CT image to group CT lung cancer region. The Adaptive Mean Shift Threshold (AMST) methodology is applied to threshold the lung cancer region for segmenting the lung cancer portion. The experimental results show that the proposed methodology is providing improved accuracy and efficiency.

Volume 12 | 05-Special Issue

Pages: 906-914

DOI: 10.5373/JARDCS/V12SP5/20201834