Zeming Huang

Zeming Huang

Master's student, University of Chinese Academy of Sciences

Curriculum Vitae

Homepage · ORCID · GitHub

EDUCATION

University of Chinese Academy of Sciences · Beijing, China
Research university of China’s national academy of sciences; ranked 54th globally and 5th in China (U.S. News 2026).
Master’s student in Electronic Information · Sep 2024 – Present
School of Electronic, Electrical and Communication Engineering; GPA: 3.82/4.00.
Sun Yat-sen University · Guangdong, China
Research university under China’s Ministry of Education; ranked 85th globally and 8th in China (U.S. News 2026).
B.Eng. in Marine Engineering and Technology · Sep 2020 – Jun 2024
GPA: 3.36/4.00.

RESEARCH INTERESTS

AI for Health, with interests in robust representation learning, medical image analysis, physiological signal modeling, and multimodal learning under incomplete annotations and limited modality availability. Particularly interested in cardiovascular applications, including ECG analysis and longitudinal modeling.

RESEARCH EXPERIENCE

Individualized ECG Waveform Forecasting · Aug 2025 – Present
University of Chinese Academy of Sciences
Clinical collaborators: Fuwai Hospital; Aerospace Center Hospital, Beijing.

  • Investigated patient-specific, year-scale ECG forecasting under sparse follow-up and heterogeneous acquisition conditions, aiming to distinguish stable individual morphology from longitudinal change and produce waveforms testable against future recordings.
  • Built ECG preprocessing and representation-learning pipelines across over one million public records and over 100,000 hospital recordings; contributed data curation, full-waveform and P/T-wave analysis, fixed-readout comparisons, and follow-up validation to baseline-conditioned and historical-record forecasting studies.
  • Baseline-conditioned forecasts showed correspondence with within-person follow-up morphology and fixed clinical readouts, including evaluation in a 959-participant clinical subset. In historical-record forecasting, CLEF achieved a normalized RMSE of 0.077 versus 0.106 for an LSTM baseline.

NOR-TL: Single-Lead Paroxysmal AF Detection · Apr 2026 – Sep 2026
University of Chinese Academy of Sciences
Clinical collaborator: Aerospace Center Hospital, Beijing.

  • Addressed limited interval-annotated ECG cohorts for intermittent AF monitoring, aiming to improve window-level recognition and time-point detection by combining large-scale recording labels with scarce temporal annotations.
  • Designed and implemented NOR-TL with a frozen normal-reference encoder, an adaptive AF stream, and separate supervision for pooled detection and sample-wise prediction. Evaluated on public, clinical-grade AFDB and external LTAFDB and CPSC2021 recordings with patient- or recording-disjoint partitions.
  • Achieved 98.0% temporal AUROC and 0.838 time-point MCC on 1,425 external CPSC2021 recordings. Ablations supported normal-reference conditioning and decoupled supervision. First-author manuscript submitted to ICASSP 2027.

MRI-Privileged Learning for Prostate Ultrasound · Sep 2024 – Present
University of Chinese Academy of Sciences
Clinical collaborator: China-Japan Friendship Hospital, Beijing.

  • Addressed unreliable prostate boundaries in noisy, low-contrast transrectal ultrasound (TRUS) and the acquisition burden of routine MRI, aiming to improve segmentation and volume measurement while keeping deployment limited to a single ultrasound frame.
  • Designed and implemented MIRAUS with utility-guided 2.5D MRI aggregation and ROI-bounded residual distillation, using MRI only during training and one TRUS frame at inference; developed complementary PRISM. Conducted patient-disjoint five-fold validation on 73 patients and automatic versus box-assisted evaluation on 24 external patients.
  • MIRAUS achieved 93.20% Dice and 3.62 mm HD95 internally; external Dice increased from 71.30% with automatic inference to 84.72% with box assistance. Model-derived calipers yielded 4.68 mL MAE against reference-caliper ellipsoid volumes in a simulated measurement workflow.

PUBLICATIONS AND MANUSCRIPTS

Accepted

  • PRISM: A Privileged Representation-Informed Synergistic Model for Transrectal Ultrasound-Only Prostate Segmentation. IEEE BIBM 2026, accepted. Co-first author (listed first).

Submitted / Under Review

  • NOR-TL: Normal-Reference Temporal Learning with Decoupled Supervision for Paroxysmal AF Detection. Submitted to ICASSP 2027. First author.
  • A Cross-Temporal Latent Evolution Framework for Year-Scale ECG Waveform Forecasting from Historical Records. Information Sciences, under review. Fourth author.

Other Manuscripts

  • MIRAUS: MRI-Informed Residual Adaptation for Single-Frame TRUS Prostate Boundary Measurement. In preparation for submission to Measurement. Co-first author (listed first).
  • Individualized ECG Waveform Forecasting Enables Testable Predictions of Longitudinal Morphology. Manuscript. Second author.
  • BioShoulder: A Large-Scale Biomechanical Dataset for Shoulder Muscle Activations. Manuscript. Third author.

ADDITIONAL RESEARCH AND ENGINEERING EXPERIENCE

Biomechanical Simulation and Physical Modeling of Humeral Motion · Apr 2025 – Present
University of Chinese Academy of Sciences
Collaborators: Peking University; Peking University People’s Hospital.

  • Constructed an ANSYS model of humeral motion with physiological movement constraints; compared simulated kinematics and forces with collaborator measurements, finding preliminary agreement in motion-force trends.
  • Designed and built a rotating physical orthopedic model for bench-scale validation of humeral motion and loading responses, integrating 3D mechanical design, fabrication, and assembly.
  • Co-authored BioShoulder, a dataset of simulated shoulder muscle activations, as part of this broader shoulder-biomechanics collaboration. Third author.

South China Sea Survey Center · Aug 2023 – Sep 2023
Ministry of Natural Resources of the People’s Republic of China
Buoy Laboratory Intern

  • Supported GPS and CTD sensor deployment and recovery, processed received observations, and uploaded field data to cloud-based systems.

Miniature Tracked ROV for Marine Observation · Jan 2022 – Dec 2022
Project Leader, University-level Innovation Project, Sun Yat-sen University

  • Led mechanical design and prototyping with SolidWorks, motion simulation, and 3D printing; assisted Arduino-based control and field testing, delivering a working underwater locomotion prototype.

HONORS AND AWARDS

  • Second-Class Academic Scholarship, University of Chinese Academy of Sciences · 2026
  • Merit Student, University of Chinese Academy of Sciences · 2026
  • Third Prize, South China Regional Contest, 12th National Marine Vehicle Design and Manufacturing Competition · 2024
  • Merit Award, Sun Yat-sen University Experimental Skills Competition · 2021
  • Outstanding Team Member, Sun Yat-sen University Admissions Outreach Association · 2021

SKILLS

  • Programming and Methods: Python, MATLAB; digital signal processing, ECG representation learning, multimodal learning, medical image analysis, and patient-disjoint evaluation.
  • Engineering: ANSYS, SolidWorks, AutoCAD; 3D printing and mechanical assembly.
  • Languages: Mandarin Chinese; English (CET-6).