DOI: 10.14704/nq.2022.20.8.NQ44818

Diagnosis of Pulmonary Nodules using MultiSize Multi-Branch 3D-CNN Architecture

Nageshbabu Dasari, B V Ramana Reddy

Abstract

Early prediction of lung cancer is very crucial now-a-days. 3D-CNN are widely used to extract dominant features from the given medical images. So, in this paper we designed a novel 3D-CNN architecture based on the size of the pulmonary nodule. In this architecture three branches S, M, and L are divided considering the diameter of the nodule as 9 mm, 15 mm, and above. All the features obtained from the three branches are concatenated to get the result. The designed model is validated with the publicly available LUNA-16 database is used. The model gives promising results with 92.6% CPM values

Keywords

Lung Cancer, Computer-aided diagnosis (CAD) technology, Computed Tomography (CT), Lung Nodule Detection, Deep Learning

Full Text

References