Lung Cancer Detection Using Machine Learning Algorithms

Abstract

This study's objective is to evaluate the degree of accuracy of several machine learning algorithms for detecting early-stage lung cancer. A thorough investigation showed that certain classifiers achieve higher precision but struggle to achieve 100% accuracy, while others show low accuracy. Inaccurate DICOM image processing leads to lower accuracy and greater implementation costs. Although various types of medical images are utilized in medical image processing, CT scans are typically favored due to their minimal noise. The best method for processing medical images, recognising and categorising lung nodules, extracting characteristics, and determining the stage of lung cancer is deep learning. Using segmentation techniques from the K Means algorithm and image processing techniques, this system originally extracted lung parts. After the segmented photos had been cleansed of any relevant features, they were classified using a variety of machine learning methods. In terms of accuracy, sensitivity, specificity, and processing times, the suggested methodologies were assessed.