.png){kind=icon}
.png){kind=icon}
Return
HC-HGCN: A Human-Computer Collaborative Diagnostic Model Based on Hypergraph Convolutional Neural Network for Pulmonary Nodule Imaging
.png){kind=icon}
OA
AI
DOI：10.1587/transfun.2025EAP1021.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Pulmonary nodule imaging diagnosis is in a leading position in the field of deep learning research, but few can really be deployed and promoted. In this study, we summarize the reasons that hinder the deployment of research results, and develop a pulmonary nodule diagnostic model using 1015 cases of CT (Computed Tomography) images and diagnostic image reports from Yantai Affiliated Hospital of Binzhou Medical University, where the LIDC-IDRI dataset was used for external testing. Our model includes three paths: a physician diagnostic path developed by extracting and statistical analysis of high-frequency terms in diagnostic image reports, an AI (Artificial Intelligence) diagnostic path developed by training CT images, and a human-computer collaborative diagnostic path developed by the hypergraph convolutional neural network (HGCN). The results show that both in the internal test set (AUC of 0.9745) and in the external test set (AUC of 0.9694), the human-computer collaborative path achieves optimal results, which confirms that our model can combine the experience of physicians with the computational power of AIto achieve more accurate and reliable diagnosis; in addition, the easy-to-access input data and the github-shared code also increase the possibility of model deployment.
Keywords:
human-computer collaboration
hypergraph convolutional neural network
pulmonary nodules
imaging diagnosis
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
0.4
Papers:
210
Citations:
1.3K
View more organizations