Jingyu Fu

Jingyu Fu

Southern University of Science and Technology

I am an undergraduate student in Biomedical Engineering (Intelligent Medical Engineering) at the Southern University of Science and Technology (SUSTech). My research interests lie in medical AI, multimodal medical imaging, point cloud-guided medical image registration, pathology foundation models, and contactless physiological sensing.

My current work focuses on clinically meaningful AI systems that connect model design with deployable medical workflows. I have worked on longitudinal lung CT registration, thyroid ultrasound video diagnosis, cytology/pathology risk stratification, smartphone-based physiological sensing, and full-stack medical review platforms.

Research keywords: Medical AI / Multimodal Medical Imaging / Medical Image Registration / Point Cloud Analysis / Pathology Foundation Models / Contactless Physiological Sensing / Clinical Decision Support

News

  • 2026.05: A first-author paper entitled “Video-level Thyroid Ultrasound Diagnosis via Detection, Self-supervised Representation Learning, and Web-based Visualization” has been released on ChinaXiv (ChinaXiv:202605.00048v1).
  • 2026: Started medical imaging research internship at Heilongjiang Tuomeng Technology Co., Ltd., working on point cloud-guided longitudinal lung CT registration for anatomically non-corresponding scenarios.
  • 2026: Pathology foundation-model manuscript revised and resubmitted to Modern Pathology (MODPATH-D-26-00161).
  • 2025.12: Received the Zhiren Dean’s Award, the highest honor of Zhiren College, SUSTech.
  • 2025.11: Received the First-Class Outstanding Student Scholarship at SUSTech.
  • 2025.11: Named Excellent Student Cadre Model at SUSTech, among 7 university-wide recipients.
  • 2025.06: Granted a Chinese invention patent as first student inventor for non-contact SpO2 monitoring based on smartphone DC calibration.
  • 2025: Paper on smartphone dual-camera blood pressure screening accepted by IEEE EMBC 2025.

Publications and Patents

  1. Q. Wei, Jingyu Fu, et al. Smartphone Dual-Camera Based Blood Pressure Screening Using Pulse Transit Time between Ear and Finger. IEEE EMBC, 2025. [DOI]

  2. Jingyu Fu (First Author), et al. Video-level Thyroid Ultrasound Diagnosis via Detection, Self-supervised Representation Learning, and Web-based Visualization. ChinaXiv, 2026. [PDF] [ChinaXiv:202605.00048v1]

  3. L. Shang, Y. Huang, X. Feng, Jingyu Fu, et al. A Pathology Foundation Model-Based Dual-Probability Framework for Risk Stratification and Clinical Triage of Cytologically Indeterminate Thyroid Nodules. Revised and resubmitted to Modern Pathology (MODPATH-D-26-00161), 2026.

  4. W. Wang, Jingyu Fu, H. Lai. 基于手机直流分量校准的非接触式血氧监测方法及系统 [Method and system for non-contact SpO2 monitoring based on smartphone DC component calibration]. Chinese invention patent, ZL 2025 1 1080860.5 / CN120585325B, filed Aug. 4, 2025, granted Nov. 11, 2025. [Google Patents] [CNIPA Public Service]

  5. Jingyu Fu, et al. Thyroid ultrasound AI-related invention patent. Patent under review.

Selected Research

Medical Imaging AI
Overall framework of the ultrasound thyroid nodule detection and classification system

AI-driven thyroid ultrasound video diagnosis

Developed a YOLOv12 + SimCLR framework for thyroid ultrasound video diagnosis under severe class imbalance. The system integrates lesion detection, representation learning, temporal aggregation, and a Flask + React clinical review workflow.

  • Key result: malignant recall up to 90.6%.
  • Current status: first-author ChinaXiv preprint (ChinaXiv:202605.00048v1); invention patent under review.
  • Links: ChinaXiv PDF.
Pathology Foundation Model
Pathology foundation model framework for thyroid FNA risk stratification

Pathology foundation model-based dual-probability framework for thyroid FNA triage

Contributed to a foundation-model-based framework for Bethesda III-IV thyroid cytology risk stratification. The project targets clinically challenging indeterminate nodules and emphasizes interpretable, high-confidence risk stratification.

  • Key result: AUC 0.948 for thyroid FNA risk stratification.
  • Current status: revised and resubmitted to Modern Pathology (MODPATH-D-26-00161).
Physiological Sensing
Pipeline for smartphone dual-camera blood pressure screening

Contactless physiological sensing with smartphones

Developed smartphone-based pipelines for non-contact blood pressure and SpO2 monitoring, including dual-camera PTT modeling and personalized AC-DC calibration for robust oxygen saturation estimation.

  • IEEE EMBC 2025 paper as second author.
  • Chinese invention patent granted as first student inventor.
  • Link: EMBC paper.
SpO2 Estimation
Smartphone PPG SpO2 calibration and estimation framework

Smartphone PPG-based SpO2 calibration

Designed a smartphone-based SpO2 estimation framework that extracts face and finger PPG signals, separates AC/DC components, and performs personalized ratio-of-ratios calibration.

  • Key contribution: AC-DC calibration strategy for more robust non-contact oxygen saturation estimation.
  • Output: Chinese invention patent granted, ZL 2025 1 1080860.5 / CN120585325B, granted Nov. 11, 2025. [Google Patents]

Computational Modeling and Competitions

Challenge Cup
PAD multimodal auxiliary diagnosis framework

PAD multimodal auxiliary diagnosis device

Participated as a core member in a lower-limb arterial obstruction diagnostic device integrating perfusion index, PATD-related parameters, and multimodal physiological signals.

  • Key result: classification accuracy above 90% in project evaluation.
  • Awards: Challenge Cup Guangdong Provincial Third Prize and Guangdong Climbing Program support.
Mathematical Modeling
MCM modeling framework

Mathematical Contest in Modeling

Built a modeling pipeline for medal prediction, analysis of the “great coach” effect, and additional insight generation using regression, random forest, uncertainty estimation, and interpretive analysis.

  • Award: Meritorious Winner in MCM/ICM.
  • Focus: predictive modeling, uncertainty estimation, and structured insight extraction.
Optimization Modeling
CUMCM modeling framework

China Undergraduate Mathematical Contest in Modeling

Designed a multi-regression optimization framework for pregnancy-related risk analysis, including correlation modeling, clustering, LASSO regression, logistic modeling, and threshold design.

  • Award: Provincial First Prize in CUMCM.
  • Focus: statistical modeling, optimization, clustering, and risk stratification.

Selected Projects

  • RingTool reproduction and ONNX deployment: Reproduced Tsinghua Pervasive HCI Lab’s RingTool benchmark toolkit for ring-based cardiovascular metrics and converted PyTorch models to ONNX for lightweight deployment. [my reproduction] [original repository]

  • ImageStyleStudio: Developed an image style-transfer / processing demo as an open-source programming project. [repository]

  • 2048 game project: Implemented a browser-based 2048 game project with interactive logic and front-end rendering. [repository]

  • Selected GitHub repositories: Additional public coding projects are available at github.com/ElizabethLydia.

Honors and Awards

  • 2025.12: Zhiren Dean’s Award, the highest honor of Zhiren College, SUSTech; only one recipient in the college.
  • 2025.11: First-Class Outstanding Student Scholarship, SUSTech; comprehensive evaluation ranking 3/259 (Top 1.2%).
  • 2025.11: Excellent Student Cadre Model, SUSTech; among 7 university-wide recipients.
  • 2025.07: Challenge Cup Guangdong Provincial Third Prize for a PAD multimodal auxiliary diagnosis project.
  • 2025.05: Mathematical Contest in Modeling, Meritorious Winner.
  • 2025.12: China Undergraduate Mathematical Contest in Modeling, Provincial First Prize.
  • 2026.04: Outstanding Volunteer, SUSTech.
  • 2024.11: Second-Class Outstanding Student Scholarship, SUSTech.

Service and Leadership

  • Chair, Academic Affairs Committee, Zhiren College, SUSTech.
  • Class Youth League Branch Secretary, SUSTech.
  • Head, Zhiren Chinese Traditional Music Ensemble; recognized as an Outstanding Student Club Leader.
  • Peer mentor / senior student advisor, Zhiren College, SUSTech.
  • Participant in SUSTech outreach and volunteer activities; recognized as an Outstanding Volunteer and an outstanding individual in alumni-school outreach.

Education

  • 2023 - Present, Southern University of Science and Technology, Department of Biomedical Engineering, Intelligent Medical Engineering.
    • Cumulative GPA: 3.84/4.0 (overall rank 13/49)
    • Academic-year GPA: 3.94/4.0 (major rank 1/49)
    • Comprehensive evaluation ranking: 3/259 (Top 1.2%)
    • Junior-year fall GPA: 3.94/4.0
    • IELTS: 7.0

Internships

  • 2026 - Present, Heilongjiang Tuomeng Technology Co., Ltd., Medical Imaging Research Intern. Working on point cloud-guided longitudinal lung CT registration for anatomically non-corresponding scenarios; representing airway, vascular, and lobar landmarks as semantic point clouds to support landmark matching, inconsistency detection, and deformation-field construction.

  • 2026, Pervasive Interaction Lab, Tsinghua University. Reproduced RingTool intelligent ring monitoring framework and completed PyTorch-to-ONNX deployment; reflected on cross-subject generalization limitations in PTT-based blood pressure prediction.

Contact