Lung Cancer Analysis and Diagnosis by Coalition of Photo Metric and Quality Metric Parameters

Lung cancer still seems to be a very common cause of death among the people all over the world. This paper proposes a method which is a coalition of two statistical methods Photo Metric and Quality Metric enacted for lung cancer analysis and diagnosis. Suggested methodology is established on the basis of identified statistical and mathematical parameters exploring the light variation and quality features of the lung image to carry out the evaluation for cancerous and non-cancerous types. The distinguishing Photo Metric parameters are Brightness, Luminance, Contrast and LC index, whereas the Quality Metric parameters are VIF, MSE, VSNR and PSNR. These eight identified parameters are applied one by one on the cancerous and noncancerous lung images using image processing technique using MATLAB to get an individual parameter statistical range through number of iterations in real time. If the test image falls in the particular calculated statistical range for cancerous lung images, then the test image is infected or else normal. But practically some of the values of the statistical range for cancerous and noncancerous lungs overlaps due to which the final decision is taken by an image classifier. The ANN is used as an image classifier to sort out this problem resulting in an efficient lung cancer diagnosis. The microscopic lung images are used to implement the proposed methodology. The performance analysis of each identified parameter is the heart of this paper. The authentication of the proposed methods is tested successfully using Standard Diagnostic Test.