Improvement of RT-DETR model for ground glass pulmonary nodule detection
Siyuan Tang, Qiangqiang Bao, Qingyu Ji, Tong Wang, Naiyu Wang, Min Yang, Yu Gu, Jinliang Zhao, Yuhan Qu, Siriguleng Wang
Abstract
Currently, pulmonary nodules detection work mostly focus on recognition and diagnosis of solid nodules. However, ground glass nodules have higher probability of malignancy, posing greater identification challenges and thus greater value for detection.
Introduction
Lung cancer is the leading cause of cancer related death worldwide, posing a significant threat to human health [1]. Early detection of lung nodules can help preventing lung cancer before progressing to malignancy [2]. Based on density, lung nodules are classified into solid and ground glass types.
Materials and methods
The dataset used in this paper consists of the publicly available LIDC-IDRI data [9] and a clinical data from a cooperating hospital (with a data privacy agreement signed). The LIDC-IDRI dataset was constructed by the National Cancer Institute (NCI) in collaboration with several healthcare organizations.
Results
To verify that the proposed improvement module has a lightweight effect while maintaining detection accuracy, we designed an ablation study. The dataset and experimental environment used are the same.
Limitations
Although experimental results above shows that the improved model in this paper has the advantages of fast inference speed and high detection accuracy, it also has some drawbacks, such as poor noise resistance and weak robustness. In the aforementioned experiments, the datasets used were preprocessed to remove noise interference from bones, trachea, and other elements in the images.
Conclusion
To quickly and accurately detect ground glass lung nodules, this paper proposes an improved RT-DETR model with the following enhancement. First, to increase the detection accuracy for nodules with blurred edges and small sizes, this paper introduces the FCGE module to optimize the backbone network.
Citation: Tang S, Bao Q, Ji Q, Wang T, Wang N, Yang M, et al. (2025) Improvement of RT-DETR model for ground glass pulmonary nodule detection. PLoS ONE 20(3): e0317114. https://doi.org/10.1371/journal.pone.0317114
Editor: Kannadhasan Suriyan, Study World College of Engineering, INDIA
Received: September 3, 2024; Accepted: December 21, 2024; Published: March 11, 2025
Copyright: © 2025 Tang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This paper is supported by Inner Mongolia Natural Science Foundation (Grant No.2024LHMS06006); Inner Mongolia Health Commission Project(Grant No.202201395); Baotou Municipal Health Science and Technology Project(Grant No.wsjkkj2022120); Inner Mongolia College Students' Innovation and Entrepreneurship Training Program Projects (Grant No.s202410130004); Inner Mongolia Natural Science Foundation (Grant No.2024MS06008); Inner Mongolia Natural Science Foundation (Grant No.2022MS06002,2024LHMS06024); Scientific Research for the Public Hospitals of Inner Mongolia Academy of Medical Sciences (Grant No. 2023GLLH0211).There was no additional external funding received for this study.
Competing interests: The authors have declared that no competing interests exist.
