Rasel Ahmed Bhuiyan
Portrait of Rasel Ahmed Bhuiyan

Rasel Ahmed Bhuiyan

Computer Vision & Machine Learning for Medical and Biometric Imaging

Ph.D. Candidate, Computer Science & Engineering, University of Notre Dame

I’m on the job market. I finish my Ph.D. in May 2027 and am looking for roles in computer vision, machine learning, and medical imaging AI, in industry, postdoc, or faculty positions. I’d be glad to connect.

Email me Resume (PDF) CV (PDF)

About

I’m a Ph.D. candidate in Computer Science and Engineering at the University of Notre Dame, advised by Prof. Adam Czajka in the Computer Vision Research Laboratory (CVRL). I build computer vision systems for medical and biometric imaging, especially for cases where data is limited or images are hard to work with. My work covers the full process: data collection, model training, evaluation, and releasing open-source tools.

My recent work includes the first working infant iris recognition system for clinical use (Best Paper Award, WACV 2025 CV4Smalls Workshop), a post-mortem iris recognition model that cut error by 68% over the prior state of the art, and a canthi detection model now used in NIST’s iris recognition pipeline. I’m currently working with foundation models and vision-language models (DINOv3, SigLIP, BioCLIP).

I have collaborated with NIST, NIMH, Michigan State University, and the Medical University of Warsaw. Before Notre Dame, I was a Lecturer in Computer Science and Engineering at Uttara University in Dhaka, where I taught Algorithms, Digital Logic Design, Machine Learning, and Artificial Intelligence.

14
peer-reviewed publications
Best Paper
WACV 2025 CV4Smalls Workshop
NIST
my model runs in NIST's iris recognition pipeline
NSF
funded research on infant iris recognition

Research

Infant Iris Recognition

Clinical iris segmentation and privacy-preserving data NSF-funded

  • Built the first operational infant iris identification system for clinical populations (EER 3%, AUC 0.99).
  • Outperformed evaluated commercial systems, some of which rejected every image in the clinical dataset.
  • Trained segmentation models with no infant labels, using infant-style augmentations of adult iris images.
  • Synthesized and released a privacy-preserving infant iris dataset.

Best Paper Award, WACV 2025 CV4Smalls Workshop

Paper Code

ColdIrisNet

Foundation-model adaptation for post-mortem iris recognition

  • Adapted a frozen DINOv3 (ViT-L/16) with ArcFace metric learning and a saliency-guided multi-task loss.
  • Designed a landmark-localization network to keep features aligned under extreme rotation.
  • Cut Equal Error Rate by 68% over the prior state of the art across 70 days of post-mortem change.
  • Showed strong cross-dataset and live-to-deceased transfer.

Iris Image Processing Toolbox

NIST IREX-vetted open-source toolkit

  • Delivered image processing modules to a public, NIST IREX-vetted iris recognition toolkit.
  • Paired learned models with geometry for canthi alignment, eyelid curve fitting, and projective gaze correction.
  • My canthi detection model runs in NIST's internal iris recognition pipeline.

Code

Iris Presentation Attack Detection

Generalization to unseen attacks

  • Trained models on 400K+ iris images covering live eyes and many attack types, including printouts, synthetic irises, Kindle displays, textured contact lenses, post-mortem eyes, and artificial eyes.
  • Reached over 95% accuracy on attack types held out from training.

Post-Mortem Interval Estimation

Multimodal fusion and regression under domain shift

  • Benchmarked deep regression models and built a multimodal NIR + RGB fusion framework.
  • Achieved 3.4-hour MAE on same-distribution data and 17.5-hour MAE on held-out subjects.
  • Cut cross-laboratory error by 40% with GAN-based synthetic augmentation.

Paper Code

Forensic Iris Image Synthesis

Conditional generative modeling for data-scarce domains

  • Built a conditional generator with latent hypersphere sampling for controllable identity synthesis.
  • Released 180K images across 18 rare conditions, passing 11 of 12 ISO/IEC 29794-6 quality checks.

Paper Code

Beyond Mortality

Post-mortem iris data collection and fairness analysis

  • Ran a demographic bias analysis across 32M+ comparisons covering age, gender, and decomposition stage.

Paper

Perceptographer

Generative AI for neuroscience (NIMH collaboration)

  • Built a hybrid GAN and diffusion image generation engine with latent-space exploration.
  • Used to visualize perceptual events caused by brain stimulation.

Abstract

Vision-Language Models for Forensic Estimation

Ongoing work

  • Building a retrieval-augmented, few-shot vision-language framework for post-mortem interval estimation from iris images.
  • Training SigLIP- and BioCLIP-based image-text models to predict PMI from NIR images and text descriptions.
  • Applying LoRA fine-tuning and attention-based multiple-instance learning to subject-disjoint benchmarks.

Publications

Full list on Google Scholar

2026

  1. Inf. Fusion

    ColdIrisNet: Biology-Guided Fusion for Forensic Iris Recognition

    Rasel Ahmed Bhuiyan, Adam Czajka

    Information Fusion (Elsevier), under review

  2. JOV

    Perceptographer: AI-Powered Framework for Controlled Navigation of High-Dimensional Stimulus Spaces

    Elia Shahbazi, Drew Nguyen, Rasel Ahmed Bhuiyan, Adam Czajka, Arash Afraz

    Journal of Vision, submitted

2025

  1. WACVW 2025

    Iris Recognition for Infants

    Rasel Ahmed Bhuiyan, Mateusz Trokielewicz, Piotr Maciejewicz, Sherri Bucher, Adam Czajka

    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, pp. 83–92

    Best Paper Award

    Paper Code

  2. WACV 2025

    Forensic Iris Image-Based Post-Mortem Interval Estimation

    Rasel Ahmed Bhuiyan, Adam Czajka

    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Tucson, Arizona

    Paper Code

  3. T-BIOM 2025

    Beyond Mortality: Advancements in Post-Mortem Iris Recognition through Data Collection and Computer-Aided Forensic Examination

    Rasel Ahmed Bhuiyan, Parisa Farmanifard, Renu Sharma, Andrey Kuehlkamp, Aidan Boyd, Patrick J. Flynn, Kevin W. Bowyer, Arun Ross, Dennis Chute, Adam Czajka

    IEEE Transactions on Biometrics, Behavior, and Identity Science (T-BIOM)

    Paper

  4. JOV 2025

    Visualizing the Unseen: Perceptographer, an AI Engine for Visualizing Brain-Stimulation-Induced Perceptual Events

    Elia Shahbazi, Drew Nguyen, Rasel Ahmed Bhuiyan, Adam Czajka, Arash Afraz

    Journal of Vision, Vol. 25, 2855 (abstract)

    Abstract

2024

  1. WACVW 2024

    Forensic Iris Image Synthesis

    Rasel Ahmed Bhuiyan, Adam Czajka

    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, pp. 1015–1023

    Paper Code

  2. JFI 2024

    Forensic Iris Recognition: A Survey

    Rasel Ahmed Bhuiyan, Adam Czajka

    Journal of Forensic Identification, vol. 74, no. 1, p. 38

    Paper

2021

  1. SIViP 2021

    Enhanced Bag-of-Words Representation for Human Activity Recognition Using Mobile Sensor Data

    Rasel Ahmed Bhuiyan, Shams Tarek, Hongda Tian

    Signal, Image and Video Processing (Springer Nature), pp. 1–8

    Paper

  2. IJMLC 2021

    A Bag-of-Words Based Feature Extraction Scheme for American Sign Language Number Recognition from Hand Gesture Images

    Rasel Ahmed Bhuiyan, A. K. Kundu, S. R. Shafi, A. Matin

    International Journal of Machine Learning and Computing, vol. 11, no. 1, pp. 85–91

2020

  1. Sensors 2020

    A Robust Feature Extraction Model for Human Activity Characterization Using 3-Axis Accelerometer and Gyroscope Data

    Rasel Ahmed Bhuiyan, N. Ahmed, Md. Amiruzzaman, M. R. Islam

    Sensors (MDPI), 20(23):6990

    Paper

  2. ICKII 2020

    Efficient Frequency Domain Feature Extraction Model Using EPS and LDA for Human Activity Recognition

    Rasel Ahmed Bhuiyan, Md. Amiruzzaman, Nadeem Ahmed, Md. Rashedul Islam

    3rd IEEE International Conference on Knowledge Innovation and Invention (ICKII), Kaohsiung, Taiwan, pp. 344–347

    Paper

2019

  1. ICIEV 2019

    A Hybrid Scheme Using PCA and ICA Based Statistical Feature for Epileptic Seizure Recognition from EEG Signal

    A. Matin, Rasel Ahmed Bhuiyan, S. R. Shafi, A. K. Kundu, M. U. Islam

    Joint 8th International Conference on Informatics, Electronics & Vision (ICIEV) and 3rd International Conference on Imaging, Vision & Pattern Recognition (IVPR)

    Paper

2018

  1. HNICEM 2018

    PCA and ICA Based Hybrid Dimension Reduction Model for Cardiac Arrhythmia Disease Diagnosis

    Md. Rashedul Islam, Rasel Ahmed Bhuiyan, Nadeem Ahmed, Md. Rajibul Islam

    IEEE 10th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment and Management (HNICEM), Baguio City, Philippines, pp. 1–7

    Paper

  2. ICFSP 2018

    Hand Gesture Feature Extraction Using Deep Convolutional Neural Network for Recognizing American Sign Language

    Md. Rashedul Islam, Ummey Kulsum Mitu, Rasel Ahmed Bhuiyan, Jungpil Shin

    4th International Conference on Frontiers of Signal Processing (ICFSP), Poitiers, France, pp. 115–119

    Paper

2017

  1. ICCIT 2017

    Reduction of Gesture Feature Dimension for Improving the Hand Gesture Recognition Performance of Numerical Sign Language

    Rasel Ahmed Bhuiyan, Abdul Kawsar Tushar, Akm Ashiquzzaman, Jungpil Shin, Md. Rashedul Islam

    20th International Conference of Computer and Information Technology (ICCIT), Dhaka, Bangladesh, pp. 1–6

    Paper

Experience

  1. Jan 2022 – Present

    Graduate Research Assistant

    Computer Vision Research Laboratory (CVRL), University of Notre Dame, Notre Dame, IN

    • Computer vision and machine learning for biometric recognition under hard conditions, including infant and post-mortem iris recognition.
    • Collaborations with NIST, NIMH, Michigan State University, and the Medical University of Warsaw.

  2. Oct 2018 – Dec 2021

    Lecturer

    Department of Computer Science and Engineering, Uttara University, Dhaka, Bangladesh

    • Taught undergraduate courses in Algorithms, Digital Logic Design, Machine Learning, and Artificial Intelligence.
    • Led research on human activity recognition and EEG seizure recognition (5 publications).

  3. Mar 2016 – Sep 2018

    Research Assistant

    Computer Vision & Pattern Recognition Lab, University of Asia Pacific, Dhaka, Bangladesh

    • Research on hand gesture and sign language recognition, and ECG arrhythmia classification.

Education

  1. Jan 2022 – May 2027 (expected)

    Ph.D., Computer Science & Engineering

    University of Notre Dame

    Dissertation focus: iris recognition at life extremes. Advisor: Adam Czajka

  2. Aug 2025

    M.Sc., Computer Science & Engineering

    University of Notre Dame

    Focus: forensic iris recognition. GPA: 3.83/4.00

  3. Mar 2018

    B.Sc., Computer Science & Engineering

    University of Asia Pacific, Dhaka

    Graduated with highest distinction. GPA: 3.94/4.00

Teaching

2022 – Present

Graduate Teaching Assistant

University of Notre Dame

  • Helped develop assignments, labs, and project materials; held office hours and graded student work.
  • Mentored two undergraduate students (2023) in building an iris presentation attack detection dataset.

Courses

  • CSE 60535 Graduate Computer Vision (Fall 2025)
  • CSE 60868 Graduate Neural Networks (Spring 2025)
  • CSE 40535 Computer Vision (Fall 2023, Fall 2024)
  • CSE 40868 Neural Networks (Spring 2023, Spring 2024)
  • CSE 20312 Data Structures (Fall 2022)
  • CSE 30332 Programming Paradigms (Spring 2022)

2018 – 2021

Lecturer (primary instructor)

Uttara University, Dhaka

  • Designed course materials, assignments, and exams; supervised undergraduate final-year projects.
  • Taught online over Zoom during COVID-19.

Courses

  • Algorithms
  • Digital Logic Design
  • Machine Learning
  • Artificial Intelligence

Honors

  • Best Paper Award, CV4Smalls Workshop, IEEE/CVF WACV 2025
  • NSF-funded research: Infant Biometric Identification (Grant 2237880)
  • Vice Chancellor’s Award for six consecutive semesters of academic distinction, University of Asia Pacific

Talks

  • Paper presentation, IEEE/CVF WACV 2025, Tucson, Arizona
  • Paper presentation, IEEE/CVF WACV 2024 Workshops, Waikoloa, Hawaii
  • Paper presentation, IEEE ICCIT 2017, Dhaka, Bangladesh

Service

  • Reviewer, IEEE/CVF WACV 2027 and 2025
  • Reviewer, International Conference on Science and Contemporary Technologies (ICSCT) 2021
  • Social Chair, Bangladeshi Student Association, University of Notre Dame (2023–2024)
  • President, CSE Student Association, University of Asia Pacific (2016–2017)
  • Departmental committees at Uttara University (routine, accreditation, faculty recruitment)