Research of the possibility of using neural networks to identify areas of interest in tomographic data

The paper is devoted to the study of the possibility of using neural networks to identify areas of interest in tomographic data. Networks of architectures InceptionV3, VGG19, and Resnet were examined in a study. As a result of the study, training and testing of the pre-trained ResNet34 network was c...

Full description

Saved in:
Bibliographic Details
Main Authors: Staroverov, N. E., Gryaznov, A. Y., Kholopova, E. D., Bessonov, V. B., Larionov, I. A.
Format: Conference Proceeding
Language:English
Subjects:
Model testing
Neural networks
Training
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The paper is devoted to the study of the possibility of using neural networks to identify areas of interest in tomographic data. Networks of architectures InceptionV3, VGG19, and Resnet were examined in a study. As a result of the study, training and testing of the pre-trained ResNet34 network was carried out. The trained model was tested on a test sample of 60 X-ray images of objects that were not used for training. The average error in the test sample was 20.48, which is a satisfactory result.
ISSN:0094-243X
1551-7616
DOI:10.1063/5.0013424