Profile Photo

Md. Atiqur Rahman

Coding simplifies life

Email

Google Scholar

Github

Hi, I’m Atiq. I am currently a Lecturer in CSE at IUT. Alongside teaching, I am passionate about research, particularly in developing efficient machine learning systems that can learn from both visual and linguistic data, adapt more like humans, and remain reliable under challenging environments. With this objective, I have explored diverse research areas, including medical image analysis, few-shot learning, federated learning, and the application of Multimodal Large Language Models.

News and Updates

  • Apr 01, 2026: Our FedOAP paper got accepetd to present in ICPR 2026
  • Sept 25, 2025: Our BdSL-SPOTER paper got accepetd to present in ISVC 2025
  • Jul 25, 2025: Attended CLNLP 2025 and received the Best Paper Award.
  • Jun 25, 2025: Our paper got accepeted to present in CLNLP 2025

Research

LLM-Based Evaluation of Low-Resource Machine Translation: A Reference-less Dialect Guided Approach with a Refined Sylheti-English Benchmark

Md Atiqur Rahman, Sabrina Islam, and Mushfiqul Haque Omi

CLNLP 2025
Paper Code Dataset

This work aims to equip LLMs with sufficient information to improve MTE in low-resource and linguistically diverse settings.

FedOAP: Cross-Organ Feature Sharing for Rapidly Adaptable Federated Tumor Segmentation

*Ishmam Tashdeed, *Md. Atiqur Rahman, Sabrina Islam and Md. Azam Hossain

ICPR 2026
Paper Code

In this study, we try to solve the problem of data scarcity due to privacy constraint exploiting Personalized Federated Learning and cross-organ knowledge.

Bdsl-spoter: A transformer-based framework for bengali sign language recognition with cultural adaptation

Sayad Ibna Azad and Md. Atiqur Rahman

ISVC 2025
Paper Code Dataset

In this work, we focus on bridging the gap between Bangla and Western Sign Language Recognition tasks through the use of an efficient pose-based transformer architecture.

FUSED-Net: Detecting Traffic Signs with Limited Data

Md. Atiqur Rahman, Nahian Ibn Asad, Md. Mushfiqul Haque Omi, Md. Bakhtiar Hasan, Sabbir Ahmed and Md. Hasanul Kabir

Under Review
Paper Code Dataset

This work proposes an algorithm that can detect traffic Signs under challenging conditions utilizing FSL.

Rep3net: An approach exploiting multimodal representation for molecular bioactivity prediction

Sabrina Islam, Md Atiqur Rahman, Md. Bakhtiar Hasan, and Md. Hasanul Kabir

Under Review
Paper Code Dataset

In this work, we try to build robust bioactivity predictor using three distinct yet complementary representations.

Projects

SpikeX

Using Spiking Neural Networks to develop an energy-efficient YOLOv4.

Few Shot Traffic Sign Classification

Utilizing metric based few shot classifiers for bench-marking GTSRB.

Save The Assassin

Designed and built dynamic 2D game from scratch using C# and Unity Game Engine.

Courses

Mobile Application and Development
Materials
Software Security
Materials
Algorithm Engineering
Materials
Ecommerce and Web Security