Methodology for Studying Neural Network Descriptors in Solving the Problem of Finding Anatomical Layers in Computed Tomography Images of the Lungs
https://doi.org/10.35596/1729-7648-2025-23-1-60-67
Abstract
The search for anatomical layers in lung CT images will simplify the task of diagnosis and treatment planning, as well as automate the process of image partitioning when preparing a training sample. The paper proposes a methodology for comparison of neural network descriptors and selection of an optimal neural network method for searching for similar anatomical regions. Neural network approaches are compared with traditional methods and a hybrid search algorithm based on the joint use of traditional and neural network methods is proposed. Using the proposed algorithm, the neural network search result for anatomical patterns, expressed in mm to the searched layer, was improved by 47 % for the first ten heart-class images found and by 18 % for images with positions from 10 to 100. The final anatomical region search result was improved over using traditional approaches by 9.7 % for retrieved images with positions from 10 to 100 and by 2 % for the first ten retrieved images.
Keywords
co-occurrence matrices,
neural network descriptors,
deep learning,
convolutional neural networks,
computed tomography images
Supporting agencies: This work was sponsored by the Belarusian Republican Foundation for Fundamental Research (project No F22KITG-001). The author expresses the gratitude to the Cand. of Sci., Associate Professor Vassili Alekseevich Kovalev and Cand. of Sci., Associate Professor Pavel Viktorovich Kamlach for valuable advice and comments when working on the article.
About the Author
A. A. Kosareva
Belarusian State University of Informatics and Radioelectronics; United Institute of Informatics Problems of the National Academy of Sciences of Belarus
Belarus
Belarus
Kosareva Aleksandra Andreevna, Assistant at the Electronic Engineering and Technology Department; Junior Researcher
220013, Minsk, P. Brovki St., 6
References
1. Willemink M. J., Koszek W. A., Hardell C., Jie Wu, Fleischmann D., Harvey H., et al. (2020) Preparing Medical Imaging Data for Machine Learning. Radiology. 295 (1), 4–15. DOI: 10.1148/radiol.2020192224.
2. Kosareva A. A., Kamlach P. V., Kovalev V. A. (2022) Determination of a Similar Anatomical Area on a Chest CT Image Using Traditional Image Feature Extraction Methods. Doklady BGUIR. 20 (5), 48–56. https://doi.org/10.35596/1729-7648-2022-20-5-48-56 (in Russian).
3. Morozov S. P., Andreychenko A. E., Pavlov N. A., Vladzymyrskyy A. V., Ledikhova N. V., Gombolevskiy V. A., et al. (2020) MosMedData: Chest CT Scans with COVID-19 Related Findings Dataset. arXiv:2005.06465.
4. Kosareva A. A., Snezhko E. V., Kamlach P. V., Kovalev V. A. (2021) Pulmonary Pathologies CT Images Archive`S Research and Preparation for a Given Anatomic Area Automatic System. BIG DATA and Advanced Analytics, Collection of Scientific Articles of the VII International Scientific and Practical Conference, Minsk, 19–20 May. Minsk, Bestprint Publ. 253–257 (in Russian).
5. Xiu Liu, Chris Aldrich (2022) Deep Learning Approaches to Image Texture Analysis in Material Processing. Metals. 12 (2). https://doi.org/10.3390/met12020355.
6. Yadav S. S., Jadhav S. M. (2019) Deep Convolutional Neural Network Based Medical Image Classification for Disease Diagnosis. Journal Big Data. (6). https://doi.org/10.1186/s40537-019-0276-2.
7. Krizhevsky A., Sutskever I., Hinton G. E. (2017) ImageNet Classification with Deep Convolutional Neural Networks. Communications of the ACM. 60 (6), 84–90. DOI: 10.1145/3065386.
8. Simonyan K., Zisserman A. (2014) Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv:1409.1556. 1.
9. He Kaiming, Zhang Xiangyu, Ren Shaoqing, Sun Jian (2015) Deep Residual Learning for Image Recognition. arXiv:1512.03385.
10. Howard A. G., Zhu M., Bo Chen, Kalenichenko D., Wang W., Weyand T. (2017) MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv:1704.04861v1.
11. Kolesnikov A., Beyer L., Zhai X., Puigcerver J., Yung J. (2019) Big Transfer (BiT): General Visual Representation Learning. arXiv:1912.11370.
12. Mingxing Tan, Quoc V. Le (2019) EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. arXiv:1905.11946.
13. Chollet F. (2017) Xception: Deep Learning with Depthwise Separable Convolutions. Conference on Computer Vision and Pattern Recognition.
14. Kosareva A. A., Kamlach P. V., Kovalev V. A, Snezhko E. V., Radzhabov A. G. (2022) Search for Similar Anatomical Areas on CT Images of Lungs Using Convolutional Neural Networks. BIG DATA and Advanced Analytics, Collection of Scientific Articles of the VIII International Scientific and Practical Conference, Minsk, 11–12 May 2022. Minsk, Belarusian State University of Informatics and Radioelectronics. 476–482 (in Russian).
Review
For citations:
Kosareva A.A. Methodology for Studying Neural Network Descriptors in Solving the Problem of Finding Anatomical Layers in Computed Tomography Images of the Lungs. Doklady BGUIR. 2025;23(1):60-67. (In Russ.) https://doi.org/10.35596/1729-7648-2025-23-1-60-67
Views: 310
Email this article 