Deyan Chen

Mr. Deyan Chen

orcid

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 August 2019 in IEEE Access
Verified
Journal Article
Open access
DCDeyan Chen
Deyan Chen
HZHong Zhao
Hong Zhao
XZXia Zhang
Xia Zhang
The health care domain is a knowledge-intensive domain. The quality of clinical diagnosis relies mainly on the medical knowledge and experience held by doctors. However, the ability of a single doctor is very limited, so the quality of clinical diagnosis is currently not high. In this paper, an aided diagnosis method based on domain semantic knowledge bases is proposed. Firstly, a domain semantic knowledge base is established by extracting and refining the knowledge of the medicine subject matter domain from the Freebase RDF dumps. Then, based on the semantic knowledge base, the algorithms for calculating the weights of the symptoms in the knowledge base, the relative weights of the diseases related to the input symptom set from a patient, and the related symptom set related to the input symptom set from the patient are proposed. Finally, the clinical medical record data of several common diseases are selected to make an evaluation on the proposed method. For each medical record, the symptom information is extracted from the chief complaint as the patient’s input symptom set. Based on the input symptom set, the method of this paper is used to obtain the list of related diseases and the ranking of disease relative weights. From the disease relevance rankings, the Top 1 (first diagnosis) and Top-3 (first 3 diagnoses) are compared with the doctor’s diagnoses in the medical records. Among them, ovarian cyst has the highest Top-1 and Top-3 hit rates of 67.3% and 89.1%, respectively. Followed by acute upper respiratory tract infection, Top-1 and Top-3 hit rates are 56.6% and 85.2%, respectively. The average Top-1 and Top-3 hit rates are 47.9% and 79.7%, respectively. Compared with the relevant methods, the method of this paper is better. The evaluation results show that based on the domain semantic knowledge base and the aided diagnosis method of diseases constructed in this paper, it is possible to provide aided diagnosis services of a large number of common diseases for general practitioners (especially inexperienced doctors) at the grass-roots level as well as self-diagnosis services of diseases for patients.

27 December 2024 in Medical Mycology
Verified
Journal Article
Closed access
YLYun Li
Yun Li
DCDeyan Chen
Deyan Chen
YZYouwen Zhang
Youwen Zhang
SLShuyi Liu
Shuyi Liu
LLLina Liang
Lina Liang
JZJinping Zheng
Jinping Zheng
In clinical practice, differentiating among pulmonary mucormycosis (PM), invasive pulmonary aspergillosis (IPA), and pulmonary tuberculosis (PTB) can be challenging. We aimed to evaluate the performance of chest CT-based artificial intelligence (AI) models in distinguishing among these three diseases. Patients with confirmed PM, IPA, or PTB were retrospectively recruited from three tertiary hospitals. Two models were developed: the unannotated supervised training (UST) model, trained with original CT images and the annotated supervised training (AST) model, trained with manually annotated lesion images. A network questionnaire with 20 cases was designed to assess the performance of clinicians. Sensitivity, specificity, and accuracy were calculated for both models and clinicians. A total of 61 PM cases, 136 IPA cases, and 155 PTB cases were included in the study. In the internal validation set, both models had an accuracy of 66.1%. The UST model had sensitivities of 27.3%, 73.9%, and 76.0% for PM, IPA, and PTB, while the AST model had sensitivities of 9.1%, 69.6%, and 88.0% for the same conditions. In the external validation set, both models had an accuracy of 57.6%. The UST model had sensitivities of 0, 85.7%, and 53.3% for PM, IPA, and PTB, respectively, while AST model had sensitivities of 0, 42.9% and 83.3%. A total of 112 clinicians had an accuracy of 42.9%, with sensitivities of 31.5%, 43.4%, and 48.0% for PM, IPA, and PTB. We demonstrated that two AI models showed comparable performance in diagnosing three diseases. Both models effectively detected IPA and PTB but struggled to identify PM.

4 September 2018 in IEEE Access
Verified
Journal Article
Open access
DCDeyan Chen
Deyan Chen
HZHong Zhao
Hong Zhao
In the process of constructing a domain semantic knowledge base based on ontologies, reusing existing domain knowledge bases not only facilitates sharing, integration, and reuse of the domain semantic knowledge base but also can accelerate the construction of the domain semantic knowledge base. The open and fast growing Freebase database is a good data source, which can be reused to construct the domain semantic knowledge base. However, extracting domain knowledge from the Freebase Resource Description Framework (RDF) dumps faces many challenges. For example, the dump package is too large to read or load; the dump package contains a lot of unnecessary and redundant facts; some ill-formed triples may cause the load to fail, and so on. In response to these obstacles and the deficiencies of existing research, this paper proposes a method to extract domain knowledge quickly, accurately, and completely from the Freebase RDF dumps and describes the domain knowledge using the semantic constructs in ontology standard description languages. Taking extracting the ontology schema and instance data of the medicine domain, including the facts pointing to semantically related domains, as an example, the principle and implementation process of the method are explained in detail and the algorithms of the key processes are described. Finally, the method of this paper is evaluated, including the comparison and analysis of related methods with work objectives, software tools used, processing results, processing performance, accuracy, completeness, and reusability.

8 December 2025 in Symmetry
Verified
Journal Article
Open access
PDF file
GMGuangkun Ma
Guangkun Ma
LKLinghui Kong
Linghui Kong
MGMo Guan
Mo Guan
YMYanhong Meng
Yanhong Meng
DCDeyan Chen
Deyan Chen
Coronary artery segmentation in CTA images remains challenging due to blurred vessel boundaries, unclear structural details, and sparse vascular distributions. To address these limitations, we propose DMF-Net (Dual-path Multi-scale Fusion Network), a novel multi-scale feature fusion architecture based on UNet++. The network incorporates three key innovations: First, a Dynamic Buffer–Bottleneck–Buffer Layer (DBBLayer) in shallow encoding stages enhances the extraction and preservation of fine vascular structures. Second, an Axial Local–global Hybrid Attention Module (ALHA) in deep encoding stages employs a dual-path mechanism to simultaneously capture vessel trajectories and small branches through integrated global and local pathways. Third, a 2.5D slice strategy improves trajectory capture by leveraging contextual information from adjacent slices. Additionally, a composite loss function combining Dice loss and binary cross-entropy jointly optimizes vascular connectivity and boundary precision. Validated on the ImageCAS dataset, DMF-Net achieves superior performance compared to state-of-the-art methods: 89.45% Dice Similarity Coefficient (DSC) (+3.67% vs. UNet++), 3.85 mm Hausdorff Distance (HD, 49.1% reduction), and 0.95 mm Average Surface Distance (ASD, 42.4% improvement). Subgroup analysis reveals particularly strong performance in clinically challenging scenarios. For small vessels (<2 mm diameter), DMF-Net achieves 85.23 ± 1.34% DSC versus 78.67 ± 1.89% for UNet++ (+6.56%, p < 0.001). At complex bifurcations, HD improves from 9.34 ± 2.15 mm to 4.67 ± 1.28 mm (50.0% reduction, p < 0.001). In low-contrast regions (HU difference < 100), boundary precision (ASD) improves from 2.15 ± 0.54 mm to 1.08 ± 0.32 mm (49.8% improvement, p < 0.001). All improvements are statistically significant (p < 0.001).

28 June 2025 in Mycoses
Verified
Journal Article
Closed access
YLYun Li
Yun Li
FHFeifei Huang
Feifei Huang
DCDeyan Chen
Deyan Chen
YZYouwen Zhang
Youwen Zhang
XZXia Zhang
Xia Zhang
JZJinping Zheng
Jinping Zheng

3 August 2026 in Biomedical Signal Processing and Control
Verified
Journal Article
Closed access
LMLijuan Ma
Lijuan Ma
DCDeyan Chen
Deyan Chen
XQXingshun Qi
Xingshun Qi
WCWei Cai
Wei Cai
YJYi Jing
Yi Jing
XZXia Zhang
Xia Zhang