Proposed methodology for Early Detection of Lung Cancer with low-dose CT Scan using Machine Learning

Lung cancer is one of the most lethal types of cancer in humans, with the highest mortality rate of any cancer. It is critical to detect lung nodules at an early stage in order to save a person’s life, because lung cancer is difficult to control and diagnose at a later stage. Researchers have utilized a variety of approaches to detect lung cancer in its early stages, but Low Dose CT scans have proven to be the most effective in identifying lung nodules. In LDCT, dose of X-ray is very low as compared to normal CT Scan, but the quality of LDCT scan image is poor as compare to Normal CT Scan. So, in this paper we purposed a methodology for detection of lung cancer in early stage with LDCT images. Our proposed methodology depends on Machine learning and Deep Learning approaches. We use CNN technique for feature selection and classification. In nodule detection and classification CNN gives very good results. Also we describe the available datasets as well as we also describe the future scope and challenges in this field. We believed that this paper gives a new direction to researchers for early detection of lung cancer with LDCT images.