Detection of Lung Cancer at Early Stage Using Neural Network Techniques for Preventing Health Care

Nowadays cancer has become huge threat in human life .There are different kinds of cancer, Lung cancer is common type of cancer causing very high ephemerality rate. Lung cancer is a serious illness which can be cured if it is diagnosed at prior stages. In this paper, we address the problem of extraction and segmentation the sputum cells based on the analysis of sputum color image with the aim to attain a high specificity rate and reduce the time consumed to analyze such sputum samples. A lung cancer risk prediction system is proposed here which will detect lung cancer at an early stage using CT scan images of DICOM format. One of the key challenges is to remove white Gaussian noise from the CT scan image, which is done using non local mean filter and to segment the lung Otsu’s thresholding is used. The textural and structural features are extracted from the processed image to form feature vector. Image Processing plays significant role in cancer detection, Various techniques used in Image Processing for information retrieval are Image acquisition, Noise Removal, Segmentation, and Morphological operations etc. In this approach, three classifiers are applied for the detection of lung cancer to find the severity of disease. It has been found from results that SVM classifier achieves higher accuracy of 95.12% while ANN classifier achieves 92.68% accuracy on the given data set and k-NN classifier shows least accuracy of 85.37%.