Page 26 - TD-2-1

Tumor Discovery

ORIGINAL RESEARCH ARTICLE
An approach for classification of lung nodules

Naveen HM *, Naveena C , and Manjunath Aradhya VN 2
1
1
1 Department of Computer Science Engineering, SJB Institute of Technology, Bangalore, Affiliated to
Visvesvaraya Technological University, Belagavi, Karnataka, India
2 Department of Master Computer Application, JSSTU, Mysuru, Affiliated to JSS Science and
Technology University, Mysuru, Karnataka, India

Abstract

The main objective of the proposed work is to develop an automated computer-
aided detection (CAD) system to classify lung nodules using various classifiers
from computed tomography (CT) images. One of the most important steps in lung
nodule detection is the classification of nodule and non-nodule patterns in CT. The
early detection of the condition helps lower the mortality rate. The developed CAD
systems consist of segmentation, feature extraction, and classification. In this work,
a filter method is used to segment the infected region. Later, we extracted features
through and fed into classifiers such as Decision Stump (DS), Random Forest (RF),
and Back Propagation Neural Network (BPNN). The experimentation was conducted
on LIDC-IDRI dataset, and the results with BPNN outperformed those with DS and RF
classifiers.

Keywords: Decision stump; Random forest; AdaBoost-Decision stump; AdaBoost-
Random forest; Back propagation neural network

*Corresponding author:
Naveen HM 1. Introduction
(naveenhm056@gmail.com)
The second most frequent cancer in both men and women is thought to be lung cancer.
Citation: Naveen HM, Naveena C, It is the main factor in cancer-related fatalities. According to the most recent estimates,
and Aradhya VNM, 2023, An
approach for classification of lung there are around 7.6 million cancer-related deaths globally each year, according to the
nodules. Tumor Discov, 2(1): 317. most recent numbers supplied by the World Health Organization . Furthermore, it is
[1]
https://doi.org/10.36922/td.317 anticipated that the number of deaths from lung cancer would keep increasing, reaching
Received: December 28, 2022 almost 17 million in 2030. Successful treatment of lung cancer depends greatly on early
Accepted: February 17, 2023 detection. Significant data suggest that early identification of lung cancer will reduce
Published Online: March 8, 2023
mortality rates . Lung cancer in an early stage manifests itself as a pulmonary nodule,
[2]
Copyright: © 2023 Author(s). which grows rapidly and later becomes a tumor. The characteristics of pulmonary
This is an Open Access article nodules are based on calcification, internal structure, sphericity, speculation, subtlety,
distributed under the terms of the
Creative Commons Attribution and texture. Nodules usually appear smaller in medical images. Hence, detection of
License, permitting distribution, pulmonary nodule is one of the most challenging tasks .
[3]
and reproduction in any medium,
provided the original work is Various imaging techniques, including radiography, computed tomography
properly cited. (CT), magnetic resonance imaging (MRI), and positron emission tomography-CT
Publisher’s Note: AccScience (PET-CT), among others, can be used to detect pulmonary nodules. Radiologists
Publishing remains neutral with face a challenging problem when trying to find lung nodules on radiographs, because
regard to jurisdictional claims in
published maps and institutional nodules present behind the rib cages are hidden and the miss rate could increase up
affiliations. to 30% [4,5] . MRI and PET-CT techniques are more expensive and time-consuming. The

Volume 2 Issue 1 (2023) 1 https://doi.org/10.36922/td.317