Sheng-Chieh (Leon) Chiu

Postdoctoral Researcher · Ph.D. Biomedical Engineering

Sovereign AI and digital twin technologies for neurological and surgical care; amyloid PET/MRI quantification and trustworthy generative neuroimaging.

Get In Touch

About

I focus on the intersection of deep learning, neuroimaging, and trustworthy AI for clinical decision support.

Sheng-Chieh (Leon) Chiu

I am a Postdoctoral Researcher at the Digital Medicine and Smart Healthcare Research Center, National Yang Ming Chiao Tung University (NYCU), focusing on sovereign AI and digital twin technologies for healthcare. I completed my Ph.D. in Biomedical Engineering at the University of Alabama at Birmingham (UAB), where I developed deep learning methods for amyloid PET quantification, MRI-assisted and MRI-less brain segmentation, simultaneous PET/MR motion correction, and large-scale neuroimaging biomarker analysis. I work at the interface of trustworthy clinical AI, neuroimaging, and real-time surgical computer vision.

Experience

Postdoctoral Researcher

Digital Medicine and Smart Healthcare Research Center, NYCU · Taipei, Taiwan
Feb 2026 – Present
  • Sovereign AI and digital twin technologies for trustworthy, locally governed clinical AI in neurology and surgery.
  • Cross-modality generative AI: synthesize amyloid PET from structural MRI for burden estimation when PET is unavailable, with safeguards against false positives in healthy populations.
  • Foundation-model framework for resting-state fMRI to learn healthy brain dynamics (self-supervised), toward clinical perturbation modeling and digital twins.
  • Real-time computer vision for DaVinci robotic gynecologic surgery: detection and promptable segmentation with visibility-aware tracking under occlusion.

Graduate Research Assistant

Department of Biomedical Engineering, UAB · Birmingham, AL, USA
Sep 2020 – Aug 2025
  • Novel image processing and deep learning for amyloid PET quantification in Alzheimer’s disease: accuracy, efficiency, and accessibility.
  • MRI-assisted and MRI-less DL segmentation pipelines; quantitative equivalence to FreeSurfer and robust diagnostics across multi-site cohorts.
  • Tracer characteristic-based co-registration for simultaneous PET/MR motion correction; PET biomarker (awSUVR) for MCI-to-AD progression with domain-specific anatomy and tracer modeling.
  • Large-scale stats: Centiloid calibration, equivalence testing, ROC, bootstrap; collaboration across nuclear medicine, radiology, neurology, and CS.

Education

Sep 2020 – Aug 2025

Ph.D. in Biomedical Engineering

The University of Alabama at Birmingham (UAB), USA

GPA: 3.85 / 4.0

Dissertation: “Optimizing Quantification in Alzheimer’s Disease with PET Imaging through Advanced Imaging and Deep Learning Techniques”

Sep 2016 – Jun 2020

B.S. in Biomedical Engineering

National Cheng Kung University (NCKU), Taiwan

GPA: 3.76 / 4.30 Upper Division: 3.86 / 4.30

Technical Skills

Deep Learning

Proficiency 90%
Python TensorFlow PyTorch

Medical Imaging

Proficiency 85%
PET MRI Co-registration Segmentation Quantification

Imaging Tools

Proficiency 80%
MATLAB 3D Slicer SPM FSL FreeSurfer MRIcroGL

Research & Publications

Published
Amyloid PET quantification with deep learning segmentation models without MRI
Chiu SC, Lin YC, McConathy J, Lin SY, Fang YHD · Mar 2026
EJNMMI Physics. doi.org/10.1186/s40658-026-00854-8
In preparation
Deep learning MR-based segmentation approach for amyloid PET quantification (LEON)
Chiu SC, Lin YC, McConathy J, Lin SY, Fang YHD · Mar 2026
Target: EJNMMI Physics
Published
Motion correction of simultaneous brain PET/MR images based on tracer uptake characteristics
Chiu SC, Perucho JA, Fang YHD · Jul 2025
EJNMMI Physics 12(1):75 · doi.org/10.1186/s40658-025-00789-6
Preprint
Prediction of MCI-to-AD progression with atrophy-weighted standard uptake value ratios of ¹⁸F-Florbetapir PET
Fang YHD, Perucho JA, Chiu SC, Lin YC, McConathy J · Mar 2023
medRxiv · doi.org/10.1101/2023.03.09.23287059

Conference posters

SNMMI · Jun 2024
Evaluation of Deep Learning Models for Brain Parcellation in Neuroimaging of Alzheimer’s Disease
Abstract link
SNMMI · Jun 2023
Correction for Involuntary Motion of Simultaneous PET/MR Brain Scans Based on Regional Tracer Characteristics
Abstract link
WMIC · Oct 2021
Deep learning–based image processing and analysis with cloud computing for open-source imaging software
Program link · IQMLServer (GitHub)

Featured projects

LEON brain segmentation

MR-based deep learning segmentation (LEON) for amyloid PET quantification—diagnostic performance and equivalence to FreeSurfer across large neuroimaging cohorts.

Python TensorFlow PET/MRI

MRI-less amyloid PET

Deep learning quantification without MRI using synthetic CT for training; equivalent to MRI-based standard and robust on external PET/CT cohorts.

PyTorch PET/CT Deep learning

TCBC motion correction

Tracer characteristic–based co-registration for simultaneous PET/MR: reduced misalignment, improved quantification and task detectability.

MATLAB PET/MR Registration

Face mask detection

Modified VGG with four face detectors to estimate mask-wearing rates in a region (course project).

Python Transfer learning Vision

EmoSpace

Mobile game for emotion-recognition training in children with ASD (ABA-inspired); RehabWeek 2019 Student Design Challenge, 3rd place (RESNA).

Unity Mobile ASD

Leadership & activities

Vice President

Taiwanese Student Association at UAB · Jul 2021 – Aug 2022

Team lead

EmoSpace project · Oct 2018 – Dec 2019

Executive secretary

Student Association, Biomedical Engineering Dept., NCKU · Sep 2018 – Jun 2019

International exhibitor

Taiwan Festival in Japan (Sendai) · Mar 2019 – Oct 2019