Personal CV Ph.D. Application for Fall 2027

2019-2024Bachelor's Degree · Clinical Medicine (Eight-Year Program)
China Medical University

2024-2027Master's Degree · Radiology and Medical Imaging
China Medical University

Presented below is an overview of my specific contributions to published and ongoing research, together with the resources I bring to the table, including computing infrastructure and access to a network of partner hospitals and institutions.

Published Research

Nature Communications, 2025

O-FGT-based neural network for microorganism recognition and motion detection

Paper: 2D (NH4)BiI3 enables non-volatile optoelectronic memories for machine learning Paper link

Optoelectronic device
Microscopic image
Motion trajectory detection

Implemented multi-object microorganism recognition based on YOLO architecture.
Simulated convolutional-layer weight deployment with synaptic weights generated from O-FGT devices, and compared O-FGT array results with GPU-based training metrics.
Linked detection results across adjacent frames using KNN to reconstruct microorganism motion trajectories.

Original Figure and Video

Nature Communications Figure 5: O-FGT-based neural network for microorganism recognition and motion detection — Fig. 5. O-FGT-based neural network for microorganism recognition and motion detection, including the short-channel O-FGT device, YOLOv8 MDR structure, O-FGT/GPU detection comparison, and trajectory tracking results.

Supplementary Movie 2. Motion trajectory tracking of microorganism detection results.

Academic Radiology, 2025

MCANet: Multimodal Cross-Attention Network for pSA-AKI prediction

Paper: Integrating Multi-Modal Imaging Features for Early Prediction of Acute Kidney Injury in Pneumonia Sepsis: A Multicenter Retrospective Study Paper link

Chest CT
Multimodal fusion
Interpretability analysis

Built the MCA-Net multimodal fusion classification model, integrating region-specific features from the lung, epicardial adipose tissue (EAT), and T4-level subcutaneous adipose tissue (T4-SAT).
Used ResNet-18 in the first stage to extract high-dimensional features from 2D CT slices of each region, followed by MSFAN-based cross-modal similarity computation and attention reweighting.
On the external test set from an independent center, the Lung + T4-SAT + EAT three-modality model achieved an accuracy of 0.981 and an AUC of 0.99.

Original Figures

Academic Radiology Figure 2: MCA-Net architecture and MSFAN module — Fig. 2. MCA-Net architecture: ResNet-18-based regional feature extraction, MSFAN cross-modal attention fusion, and ResNet-101 classification.

Academic Radiology Figure 4: deep features, clinical correlations, and Grad-CAM maps — Fig. 4. Correlations between deep features and clinical variables, with Grad-CAM attention maps for the lung, EAT, and T4-SAT regions.

Ongoing Work

Adipose-Prior-Guided CT-to-PET Cross-Modality Synthesis for Clinically Accessible Metabolic Assessment

A two-stage adipose-prior-guided framework for synthesizing whole-body 18F-FDG PET from non-contrast CT; GitHub: github.com/llj0621/Adipose_Prior_CT2PET.

CT-to-PET synthesis
3D conditional GAN
Adipose priors
Metabolic risk analysis

Ongoing work Figure 1: two-stage adipose-prior-guided CT-to-PET synthesis framework

Synthesizing Dynamic CT Perfusion Sequences via Conditional Flow Matching

A latent conditional flow matching framework generates 25-frame dynamic CTP sequences from a single baseline image while preserving perfusion hemodynamics; GitHub: to be added.

CT perfusion synthesis
Conditional flow matching
Spatiotemporal VAE
Perfusion hemodynamics

Summary figure for dynamic CT perfusion sequence synthesis from a single frame

Resources and Collaborations

Additional computing resource photo to be added

Texture to be added

Institution	Type	Collaboration Focus
Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences	Materials science research institute	Interdisciplinary device-AI research and optoelectronic computing-related collaboration.
Shengjing Hospital of China Medical University, Liaoning, China	University-affiliated hospital	Multicenter clinical imaging collaboration.
The Second Affiliated Hospital of Dalian Medical University, Liaoning, China	University-affiliated hospital	Multicenter clinical imaging collaboration.
Xing'an League People's Hospital, Inner Mongolia, China	Regional hospital	Regional clinical cohort collaboration and multicenter validation scenarios.
Haicheng Orthopedic Hospital, Liaoning, China	Regional Specialty Hospital	Specialty clinical collaboration with musculoskeletal imaging.
Northeastern University, China	Research University	Interdisciplinary AI research collaboration.
University of Lübeck, Germany	Research University	Interdisciplinary AI research collaboration.
Heke Cultural Technology (Shenyang) Co., Ltd.	Startup Company	Computing resource support and AI technology translation collaboration.