Analysis of Lung Nodule Classification with Feature Extraction.
In this paper, the four types of lung nodules are classified, i.e. well circumscribed, vascularised, juxtapleural and pleural tail in low dose computed tomography (LDCT) scan. This classifier analyses both lung nodule and surrounding anatomical structures. Also, it consists of three main stages as follows: (1) Multi level concentric partition through patch based image representation, (2) Feature set design for patch description of image, and (3) SVM classifier compute the classification probability based on level nodule and pLSA calculate the classification probability based on level context. The proposed method was evaluated on a publicly available dataset and clearly demonstrated promising classification performance.