Mashhad, Iran · Available for research collaboration

Benyamin
Gheiji

Medical student & AI / medical imaging researcher, focused on uncertainty quantification.

I build deep learning models for radiology — segmentation, registration, classification, and vision–language report generation — and work on making them report how confident they really are.

Get in touch Google Scholar GitHub LinkedIn
Portrait of Benyamin Gheiji
Open to collaboration
0Publications
& preprints
0Honors
& awards
0International
research prizes
0MD candidate
since
01 / About

Medical imaging AI, built to be reliable.

I'm a medical student and machine learning researcher working on AI for medical imaging. My work spans segmentation, registration, classification, and vision–language report generation across radiology — with a particular focus on uncertainty quantification, so that a model's confidence is calibrated, interpretable, and safe to act on in clinical practice.

  • Imaging models end to end. Deep learning on MRI and CT — segmentation, registration, classification, detection, and report drafting with vision–language models.
  • Uncertainty quantification. Calibration, Bayesian deep learning, ensembles, and conformal prediction — so a confident wrong answer never reaches the clinician unflagged.
  • Open infrastructure. Building NeuroAIHub so radiology imaging datasets are discoverable, curated, and reusable.
A radiologist reading an AI result that reports its own uncertainty A clinician sits at a reading workstation. The screen shows a chest radiograph with a highlighted nodule beside an AI assessment: a possible finding, a wide and low confidence range rather than a single number, and a banner advising a second read because the model is uncertain. AI ASSESSMENT Possible finding 34–71% CONFIDENCE UNCERTAIN — SECOND READ ADVISED
Instead of one confident number, the model reports a range — and says when a case deserves a closer look.
02 / Research

Selected publications

Peer-reviewed articles and preprints across medical imaging AI — radiogenomics, segmentation and registration, vision–language report generation, and imaging data infrastructure — with a recurring focus on uncertainty quantification.

Uncertainty quantification in medical imaging

2026
CONReg: Uncertainty-Aware Medical Image Registration Using Conformal Prediction
Gheiji B, Elyassirad D, Vatanparast M, Ahmadzadeh AM, Faghani S, Tavakoli M
Journal of Imaging Informatics in Medicine
2026
CONRep: Uncertainty-Aware Vision-Language Report Drafting Using Conformal Prediction
Elyassirad D, Gheiji B, Vatanparast M, Ahmadzadeh AM, Agah SA, Moassefi M, Tavakoli M, Faghani S
arXiv preprint
2025
CONSeg: Voxelwise Uncertainty Quantification for Glioma Segmentation Using Conformal Prediction
Elyassirad D, Gheiji B, Vatanparast M, Ahmadzadeh AM, Faghani S
American Journal of Neuroradiology
2025
Predicting IDH Mutation in Glioma Patients Using Deep Learning Algorithms with Conformal Prediction
Elyassirad D, Gheiji B, Vatanparast M, Ahmadzadeh AM, Masoudi M, Gholami M, Moassefi M, Faghani S
Journal of Imaging Informatics in Medicine

Imaging data science & AI infrastructure

2026
NeuroAIHub: An AI-Driven Framework for Automated Curation and Discovery of Neuroradiology Datasets
Gheiji B, Moradi S, Vatanparast M, Shokrani S, Bagherzadeh MA, Elyassirad D, Akinmuleya OI, Goodarzi S, Heidari-Foroozan M, Moassefi M, Rudie JD, Calabrese E, Jain R, Faghani S
American Journal of Neuroradiology

Radiogenomics & explainable AI

2026
UPhAIR: A Hybrid Pipeline for Generating Understandable Post-hoc AI Reports in Glioma IDH Mutation Status Prediction
Gorji A, Shahverdi H, Saberi A, Gheiji B, Farahani S, Azemi G, Di Ieva A
medRxiv preprint
2026
Radiomics and Deep Learning Models for Predicting Glioma p53 Status: A Diagnostic Accuracy Systematic Review and Meta-Analysis of MRI Studies
Ahmadzadeh AM, Ashoobi MA, Lomer NB, Vatanparast M, Gheiji B, Elyassirad D, Faghani S
Clinical Imaging
2025
MRI-Derived Deep Learning Models for Predicting 1p/19q Codeletion Status in Glioma Patients: A Systematic Review and Meta-Analysis
Ahmadzadeh AM, Broomand Lomer N, Ashoobi MA, Elyassirad D, Gheiji B, Vatanparast M, Rostami A, Abouei Mehrizi MA, Tabari A, Bathla G, Faghani S
Neuroradiology
2025
Improving Radiomics-Based IDH1 Prediction in Glioma Patients Using Semi-Supervised Machine Learning Models
Ahmadzadeh AM, Jafarnezhad A, Elyassirad D, Vatanparast M, Gheiji B, Faghani S
BMC Medical Imaging
2024
Comparative Analysis of 2D and 3D ResNet Architectures for IDH and MGMT Mutation Detection in Glioma Patients
Elyassirad D, Gheiji B, Vatanparast M, Ahmadzadeh AM, Kamandi N, Soleimanian A, Salehi S, Faghani S
arXiv preprint

Hemorrhage & trauma imaging AI

2025
Predicting Epidural Hematoma Expansion in Traumatic Brain Injury: A Machine Learning Approach
Hasanpour M, Elyassirad D, Gheiji B, Vatanparast M, Keykhosravi E, Shafiei M, Daneshkhah S, Fayyazi A, Faghani S
The Neuroradiology Journal
2025
Application of Deep Learning for Predicting Hematoma Expansion in Intracerebral Hemorrhage Using CT Scans: A Systematic Review and Meta-Analysis
Ahmadzadeh AM, Ashoobi MA, Broomand Lomer N, Elyassirad D, Gheiji B, Vatanparast M, Bathla G, Tu L
La Radiologia Medica
Full publication list on Google Scholar
03 / Honors

Awards & recognition

  1. 2026 $1,250 prize

    Winner of the 2026 SIIM Helen and Paul Chang Foundation New Investigator Award (USA) for our work on CONRep: Uncertainty-Aware Vision-Language Report Drafting Using Conformal Prediction.

  2. 2025 $1,000 prize

    Winner of the 2025 John L. Ulmer Award at ASFNR (USA) for our work on CONSeg: Voxelwise Uncertainty Quantification for Glioma Segmentation Using Conformal Prediction.

  3. 2024

    1st Rank among presented abstracts in the Artificial Intelligence and Data Science Panel at the 25th Annual Congress of Research and Technology of Iranian Medical Sciences Students.

  4. 2024

    Recognized among the top-ranked abstracts at the AIH Congress, held by Sharif University of Technology and Iran University of Medical Sciences.

  5. 2023

    1st Rank, Artificial Intelligence in Medical Sciences Idea Challenge — Final Round, organized by the Student Research Committee, Mashhad University of Medical Sciences.

  6. 2023

    Bronze Medalist, 15th National Entrepreneurship Olympiad for Iranian Medical Sciences Students, Iran.

04 / Open source

Code & datasets

NeuroAIHub-Registry/NeuroAIHub

The dataset registry and code behind NeuroAIHub — an AI-driven framework for the automated curation and discovery of radiology imaging datasets.

More on GitHub

05 / Background

Education, teaching & service

2022 — present
Doctor of Medicine (MD)
Mashhad University of Medical Sciences, Iran
2025
Workshop Instructor — AI Fundamentals in Healthcare
Student Research Committee, MUMS · Designed and delivered an AI fundamentals workshop series for medical students.
2023
Research Mentor
Student Research Committee, MUMS · Mentored students in research methodology and academic workflow.
Ongoing
Peer Reviewer
PLOS ONE
06 / Toolkit

Technical skills

Programming

Python — advancedSQL — intermediate

Machine learning & deep learning

PyTorchTensorFlowScikit-learn NumPyPandasMatplotlib CNNsVision TransformersLLMsVLMs SegmentationRegistrationDetectionImage Generation

Medical imaging

MRI / CT PreprocessingMONAINIfTI / DICOM RegistrationAugmentation Radiology WorkflowsRadiomics

Uncertainty quantification

CalibrationBayesian Deep Learning Variational InferenceMC Dropout Deep EnsemblesEvidential DL Conformal Prediction

Agentic AI & LLM applications

LangChainLangGraphRAG Tool UseMulti-Step Reasoning Prompt EngineeringLLM Evaluation FAISSChromaSemantic Search

MLOps & deployment

MLflowWeights & BiasesAirflow DockerTerraformGrafana FastAPIFlaskGit / GitHubCI/CD

Cloud & computing — familiar

AWSGCPAzure ML

07 / Contact

Let's build medical imaging AI worth trusting.

I'm always glad to hear about collaborations in medical imaging and radiology AI, uncertainty quantification, and open imaging infrastructure.

benyamingheiji@gmail.com BenTalks on Telegram