Marawan Elbatel

PhD Candidate in Electronic and Computer Engineering, HKUST

I develop clinically grounded multimodal agents for evidence-scarce medicine, enabling reliable diagnosis from low-cost, contrast-free imaging through data-efficient learning across institutions and countries.

My work spans medical vision-language models, federated and data-efficient learning, clinical benchmarks, and external validation, with a particular focus on translation with African healthcare institutions.

I am advised by Prof. Xiaomeng Li and serve as Vice President of SIG-AFRICAI within the MICCAI Society.

Email/CV/Google Scholar/GitHub/LinkedIn

News

2026OpenEcho accepted for an oral presentation at ESC Digital & AI Summit 2026.
2026TriCECT early accepted at MICCAI 2026 as a Spotlight paper (top 4%).
2026MedSapiens published in Medical Image Analysis.
2026Leading the AFRICAI Workshop at MICCAI 2026 and the TriALS-Report Challenge at MICCAI 2027.
2025Appointed Vice President of SIG-AFRICAI, MICCAI Society.

Selected publications

Full list on Google Scholar

OpenEcho LLMInvoke tool callsIntegrate tools' output

OpenEcho: An Agentic Framework for Echocardiography Interpretation

Marawan Elbatel and collaborators

ESC Digital & AI Summit, 2026 Oral · under embargo

An agentic echocardiography framework that plans specialist tool calls and integrates their outputs into a structured clinical pre-read.

Teaser only

TriCECT non-contrast CT to contrast-informed report overview

TriCECT: A Multi-Center Benchmark for Abdominal Disease Diagnosis and Report Generation from Non-Contrast CT

Mariam Elbakry, Aliaa Sayed Sheha, Salma Hassan Tantawy, Aya Yassin, Concetto Spampinato, Karim Lekadir, Xiaomeng Li, and Marawan Elbatel

MICCAI, 2026 Spotlight · top 4% · senior & corresponding author · mentee-led

A 3,000-patient benchmark with 12,000 triphasic CT volumes and paired reports, studying whether one non-contrast CT can recover diagnostically useful evidence normally derived from multiphasic imaging.

Paper

TriALS four-phase lesion supervision teaser

TriALS: Triphasic-Aided Liver Lesion Segmentation Benchmark in Non-Contrast CT

Marawan Elbatel and collaborators

Manuscript under submission Open benchmark

A 150-patient, 600-volume benchmark separating lesions visible on non-contrast CT from supervision recovered through arterial, portal-venous, and delayed phases.

Project & data

MedSapiens method and applications overview

MedSapiens: Taking a Pose to Rethink Medical Imaging Landmark Detection

Marawan Elbatel*, Anbang Wang*, Keyuan Liu, Kaouther Mouheb, Enrique Almar-Munoz, Lizhuo Lin, Yanqi Yang, Karim Lekadir, and Xiaomeng Li

Medical Image Analysis, 2026

A human-centric foundation model for landmark detection across anatomies and imaging modalities.

Paper / Code

GeoSapiens landmark detection architecture

Geometric-Guided Few-Shot Dental Landmark Detection with Human-Centric Foundation Model

Anbang Wang*, Marawan Elbatel*, Keyuan Liu, Lizhuo Lin, Meng Lan, Yanqi Yang, and Xiaomeng Li

MICCAI, 2025

Few-shot dental landmark detection guided by geometric structure and a human-centric foundation model.

Paper / Code

Adaptive fetal ultrasound frame selection examples

Adaptive Frame Selection for Gestational Age Estimation from Blind Sweep Fetal Ultrasound Videos

Tanya Akumu, Marawan Elbatel, Victor M. Campello, Richard Osuala, Carlos Martin-Isla, Ignacio Valenzuela, Xiaomeng Li, Bishesh Khanal, and Karim Lekadir

MICCAI, 2025 Oral · top 2.2%

Frame selection for gestational-age estimation from blind-sweep ultrasound acquired in resource-constrained settings.

Paper / Code

Federated semi-supervised learning method overview

Learning Unlabeled Clients Divergence for Federated Semi-Supervised Learning via Anchor Model Aggregation

Marawan Elbatel, Hualiang Wang, Jixiang Chen, Hao Wang, and Xiaomeng Li

Transactions on Machine Learning Research, 2024

Validation-free aggregation that estimates how useful unlabeled clients are to federated semi-supervised learning.

Paper

FD-SOS open-set dental diagnosis overview

Vision-Language Open-Set Detectors for Bone Fenestration and Dehiscence Detection from Intraoral Images

Marawan Elbatel, Keyuan Liu, Yanqi Yang, and Xiaomeng Li

MICCAI, 2024 Oral · top 2.8%

Open-set vision-language detection for localizing teeth and diagnosing bone fenestration and dehiscence.

Paper / Code

An Organism Starts with a Single Pix-Cell: A Neural Cellular Diffusion for High-Resolution Image Synthesis

Marawan Elbatel, Konstantinos Kamnitsas, and Xiaomeng Li

MICCAI, 2024

A compact neural cellular diffusion model for high-resolution medical-image synthesis and downstream augmentation.

Paper / Code

FoPro-KD Fourier-prompted knowledge distillation framework

FoPro-KD: Fourier Prompted Effective Knowledge Distillation for Long-Tailed Medical Image Recognition

Marawan Elbatel, Robert Martí, and Xiaomeng Li

IEEE Transactions on Medical Imaging, 2024

Fourier-prompted knowledge distillation for robust medical image recognition under long-tailed class imbalance.

Paper / Code

* Equal contribution.

Mentoring & academic service

Service

Career Advancement Chair, MICCAI 2028 (forthcoming)
Lead Organizer, TriALS-Report Challenge, MICCAI 2027 (forthcoming)
Vice President, SIG-AFRICAI, MICCAI Society
Lead Organizer, AFRICAI Workshop, MICCAI 2026
Lead Organizer, TriALS Challenge, MICCAI 2024–2025
Reviewer, IEEE TMI, TPAMI, TNNLS, and Medical Image Analysis

Mentoring

I mentor undergraduate and master's researchers across Hong Kong, Egypt, Zambia, and Ghana, supporting them in defining clinical questions, building collaborations, and leading first-author papers.

Recent mentees include Mariam Elbakry, Anbang Wang, Kangwa E. Mukuka, and Toufiq Musah.

Selected recognition: 2× Best Paper Awards · 3× International Medical-Imaging Challenge Wins — MICCAI DART 2022 and DeCaF 2023 Best Paper Awards; winners of the Breast FL Density 2022, EndoVis-SynISS 2023, and PS-FH-AoP 2023 challenges.