Detection of Lung Cancer using Multi Model Image Fusion Technique for CT and PET Images

Abstract

Lung cancer is a disease of abnormal multiplication of cells that grow into a tumor. The mortality rate of lung cancer is the highest among all other types of cancer. Lung cancer is one of the most serious cancers in the world, with the smallest survival rate after the diagnosis, with a gradual increase in the number of deaths every year. As such, the detection of lung cancer has been a tedious task in medical image analysis over the past few decades. If the lung cancer can be identified accurately at an early stage, the survival rate of the patients can be increased by a significant percentage. Lung as it is difficult to identify lung nodules using raw chest x-ray images, analysis of such medical images has become a tedious and complicated task. Presently, Computer Tomography (CT) images are said to be more effective in detecting and diagnosing the lung cancer. In CT scan, the images reflect the anatomical structure of bone tissues clearly and also provides better information about denser tissue. In the first part of work segmentation of lungs, a tumor in CT image is used to Spatially Weighted Fuzzy C Means Clustering (SWFCM) techniques. The overall accuracy, Sensitivity and, Predictive values achieved are 86.082%, 85.636 % and 92.673 % respectively. newlineIn Positron Emission Tomography (PET), images reflect the pathological structure of the tissue information clearly. So in the second part of the work segmentation of lungs, a tumor in PET image is used to Fuzzy Local Information C Means Clustering (FLICM) techniques. The SWFCM method enhances newlinevi newlineinsensitiveness to noise to some extent. However, in the detected tumor cluster, some small extra regions are found due to the overlapped intensity variations in the tumor region. This region affects the sensitivity determination which is considered in Fuzzy Local Information C-Mean (FLICM) method. The overall accuracy, Sensitivity and, Predictive values achieved are 89.31%, 87.27% and 95.88 % respectively. In order to strengthen the diagnosis for mass screening, the CT images and the PET images have to be fused effectively. newlineOne of the major research fields in image processing is image fusion. Image fusion is a process of combining the relevant information from a set of images, into a single image, where in the resultant fused image obtained will have more complete information of the all the input images in a single image itself. Image fusion is defined as the integration of data and information from all input registered images without any loss of information and distortion. It is not possible to get an image with all relevant features in focus, so to get all the features in one image is by fusing images with different focus settings. The applications of image fusion are considered in fields about medical imaging, military applications, commercial applications and satellite images. The proposed system used the wavelet transform to fuse CT and PET images. In this work, two fusion rules are used namely maximum absolute rule and Principle component analysis. The performance analyses of maximum absolute fusion rule for Entropy, Peak Signal to Noise Ratio (PSNR), Standard Deviation (SD), Structural Similarity Index Measure (SIM) and Root Mean Square Error (RMSE) are 5.832, newlinevii newline25.96, 34.1321, 0.421, and 3.256 respectively. In PCA (Principal Component Analysis) fusion rule Entropy, PSNR, SD, SIM and RMSE are 5.956, 26.75, 36.888, 0.432 and 3.164 respectively. The drawback of the wavelet transform is the low spectral resolution. newlineTo overcome the drawback in the wavelet transform in the next stage, Fast Discrete Curvelet Transform is proposed. The two fusion rules maximum absolute and Principal Component Analysis are used. The performance analyses of the proposed system for Entropy, PSNR, SD, SIM and RMSE are 6.328 28.47, 54.1765, 0.521 and 2.527 respectively. In Principal Component Analysis (PCA) fusion rule Entropy, PSNR, SD, SIM and RMSE are 6.786, 29.56, 56.192, 0.556 and 2.496 respectively. newlineTo overcome the drawback in Curvelet transform in next stage, Non-Subsampling Contourlet Transform is proposed. The two fusion rules maximum absolute and PCA are used. The performance analyses of maximum absolute fusion rule for Entropy, (PSNR), (SD), Structural (SIM) and (RMSE) are 6.543 22.78, 55.682, 0.435 and 2.478 respectively. In PCA fusion rule Entropy, PSNR, SD, SIM and RMSE are 7.498, 23.16, 56.182, 0.479 and 1.982 respectively. newlineTo overcome the drawbacks of the wavelet transform, Curvelet transforms and NSCT in the next stage, multimodal fusion is proposed. In this techniques, SWT and NSCT transform are used and also two fusion rules maximum absolute and PCA are newlineviii newlineused. In the first stage, SWT and PCA fusion rules are used and in the second stage NSCT and maximum absolute are used. The performance analyses of the proposed system for Entropy, (PSNR), (SD), Structural (SIM) and (RMSE) are 7.568, 49.58, 60.282, 0.623 and 0.978 respectively. To enhance the accuracy in detection of lungs tumor, higher resolution of CT and PET images may be used for future work to increase the level of accuracy. newline