Computer-Aided Detection of Human Lung Nodules on Computer Tomography Images via Novel Optimized Techniques.

BACKGROUND Globally, the most general reason for huge number of passings is Lung disease. The lung malignancy is the most shocking amongst the tumor types and it plays a significant role for the increase of death rate. It is assessed that nearly 1.2 million persons are determined to have this illness and about 1.1 million individuals are losing their lives due to this sickness in every year. The survival rate is superior if the growth is recognized at earlier periods. The premature identification of lung malignant growth isn't a simple task. Various imaging algorithms are available for detecting the lung cancer. AIM Computer aided diagnosis scheme is more useful for radiologist in detecting and identifying irregularities in advance and more rapidly. The CAD systems usually focus on identifying and detecting the lung nodules. Staging the lung cancer at its detection need to be focused as the treatment is based on the stage of the cancer. The major drawbacks of existing CAD systems are less accuracy in segmenting the nodule and staging the lung cancer. OBJECTIVE The most important intention of this work is to divide the lung nodule from CT image and classify as tumorous cells in order to identify the cancer's position with greater sensitivity, precision, and accuracy than other strategies. METHODS The primary role is defined as follows (i) for de-noising and edge sharpening of lung image, the curvelet transform is used. (ii) The Fuzzy thresholding technique is used to perform lung image binarization and lung boundary corrections. (iii) Segmentation is performed by using K-means algorithm. (iv) By using convolutional neural network (CNN), different stages of lung nodules such as benign and malignant are identified. RESULTS The proposed classifier achieves a 97.3 percent accuracy. The proposed approach is helpful in detecting lung cancer in its early stages. The proposed classifier achieved a sensitivity of 98.6 percent and a specificity of 96.1 percent. CONCLUSION The results demonstrated that the established algorithms can be used to assist a radiologist in classifying lung images into various stages, thus supporting the radiologist in decision making.