Evolution of Lung CT Image Dataset and Detection of Disease

Abstract

This paper presents CT scan image analysis, creation of database and evolution of content-based image retrieval technique for distinguishing lung cancer at early stages. The data are collected from the clinical environment and LIDC dataset. The features such as correlation, dissimilarity, cluster prominence and cluster shade are extracted at different orientations using GLCM features in the MATLAB environment and stored in the database as a trained phase. The testing image features are extracted and are analogized with the trained dataset and the appropriate out-turn is obtained. Minimum distance classifier is used to predict the clinical condition of the lung by matching the testing image and trained image.