Benign and Malignant Classification Model of Pulmonary Nodules Based on Residual Neural Network

DOI

10.2991/acsr.k.191223.038How to use a DOI?

Keywords

deep learning, residual network, pulmonary nodules, benign and malignant classification

Abstract

Computer-assisted diagnosis is of significance in the timely treatment of lung cancer with classifying benign and malignant pulmonary nodules. Aiming at improving the low accuracy rate of benign and malignant pulmonary nodules and reducing the misdiagnosis rate and wrong-diagnosis rate in computer-aided diagnosis system, a classification model of pulmonary nodules based on residual network was proposed. Firstly, selected some lung CT images from LIDC-IDRI as a data set, amplified the data by horizontal flipping, and then converted them into single channel images. After cropping and normalization, the data was finally divided into training set and test set (7:3), and used to train and test a residual network (ResNet-26). After training, test results represent that the model accuracy rate, sensitivity and specificity are 97.53%, 97.91% and 97.18%. By comparing various methods, the raised method performs better than others according to accuracy, sensitivity and specificity, which demonstrates that it has the ability to help doctors in diagnosis.

Copyright

© 2019, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhenzhe Lin
AU  - Guitang Wang
AU  - Qinshen Fu
AU  - Guozhen Wang
PY  - 2019
DA  - 2019/12/24
TI  - Benign and Malignant Classification Model of Pulmonary Nodules Based on Residual Neural Network
BT  - Proceedings of the 2019 International Conference on Big Data, Electronics and Communication Engineering (BDECE 2019)
PB  - Atlantis Press
SP  - 164
EP  - 167
SN  - 2352-538X
UR  - https://doi.org/10.2991/acsr.k.191223.038
DO  - 10.2991/acsr.k.191223.038
ID  - Lin2019
ER  -