Performance Comparison of Convolutional Neural Network (CNN) with Traditional Methods for Cancer Cell Detection
In the word cancer is one of the most dangerous diseases. The Lung cancer a type of cancer is the leading cause of death from cancer. Early detection and diagnosis of the disease is vital to prolong the lives of patients affected by the disaster. Lung cancer is detected in early-stage by using several image processing and machine learning techniques. For medical images, the Artificial Neural Network is proven to the best technique especially in the lung cancer diagnosis. Computed tomography (CT) scanned images for lung nodule detection is preferred by so many researchers. The system use CT scans images that are extracted from the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) dataset. In this paper, a new method for improving and classifying CT scan datasets of lung images based on antagonistic networks is proposed. The system proposes a Convolutional neural network (CNN) for detecting and classifying lung nodules. The proposed method makes it possible to increase the accuracy of determining the number of lungs nodules in the pulmonary area, and also allows differentiating benign and malignant nodules using the CNN architecture. Finally, CNN is shown to significantly transform early diagnosis and treatment of lung cancer. The paper also shows the performance comparison between CNN and Naïve Bayes.