Summary
I have completed PhD in Computer Science from The University of Texas at Arlington in 2023. My research
interests are in Machine Learning, Computer Vision, Deep Learning, and
Medical Image Analysis. I did my PhD under the supervision of Dr. Junzhou Huang. My research projects include Nuclei/Cell Segmentation, Detection,
Classification, etc.
I worked as a Software Engineer for two years at Therap Services. I completed my Bachelor in Computer Science & Engineering from Bangladesh University of Engineering and Technology (BUET) in 2014.
Skills
Programming :
C/C++, Java, Python, MATLAB, Scala, Prolog, Assembly 8086, Linux Shell script
ML tools :
PyTorch, TensorFlow, Keras, MatConvNet, SageMaker, TensorBoard, scikit-learn, NumPy, pandas, Matplotlib
Web :
PHP, JSP, FTL, JavaScript, HTML, CSS, XSL
Database :
MySQL, Oracle, PL/SQL
Frameworks & API :
Spring, Servlet, Hibernate, Zend, JUnit, jQuery, Bootstrap, OpenGL
Tools :
Git, AWS, Gradle, Oracle SQL Developer, LaTeX, Packet Tracer
Experiences
As a Graduate Research Assistant, I work on nuclei/cell segmentation, detection and classification. My research
works focus on developing robust Machine Learning and Deep Learning models for improving the state of the arts.
As a Graduate Teaching Assistant, my responsibilities included conducting classes, creating assignments, grading
exams and assignments, exam proctoring, and helping students with their assignments and codes.
Worked on developing new features and enhancing functionality of existing modules, proper integration
of third-party software, solving critical issues, fixing bugs, code refactoring, code reviewing, and
doing research on user authentication and access control. Technologies include Java, Oracle, Spring Framework,
Servlet, JSP, FTL, JavaScript, etc.
Projects
Domain Adaptation for Cell Segmentation:
A semantic segmentation model implemented in Python, PyTorch and Keras. Adversarial learning was applied due to the lack of annotations.
Transformers for Cell Segmentation:
Visual Transformer (ViT) based semantic segmentation framework implemented in Python and PyTorch. Self-supervised learning is used to learn the quantity of cells, which in turn leads to better segmentation performances.
Sub-type Cell Segmentation:
Simultaneous cell/nuclei segmentation and classification framework implemented in Python and PyTorch. Instance segmentation, fine-grained classification and self-supervised learning techniques were applied.
Face Recognition with Deep Convolutional Neural Network:
An implementation of Deep CNN model in which FaceNet was used as feature extractor. The model was written in Python and TensorFlow.
Cell Detection and Classification with Convolutional Neural Network:
Simultaneous cell detection and classification framework implemented in MATLAB and MatConvNet.
Deep Neural Network for Image Classification:
A deep L-layer image classifier implemented in Python. NumPy, Matplotlib, PIL and SciPy packages were used.
Traffic Crash and Driver Distractions Classification:
A crash-severity predictive model using Logistic Regression and Decision Tree algorithm. The model can also classify if a driver is distracted or not. The model was implemented in Python using Pandas and NumPy packages.
Face Recognition with Kernel Support Vector Machine:
This framework was written in MATLAB. Principal Component Analysis and Linear Discriminant Analysis were applied for dimensionality reduction. Polynomial kernel with degree 2 was chosen as the kernel. Linear SVM and K-NN algorithm were also implemented for performance comparisons.
Altered Fingerprint Matching:
An altered/obfuscated fingerprint matching framework using Support Vector Machine with radial basis kernel, and K-nearest neighbor algorithm. Model was implemented with MATLAB and libSVM library. NBIS was used for extracting features from fingerprint images.
Text article Classification with Bayesian Learning:
An implementation of Naive Bayes classifier for text documents classification. The model was written in Java.
An N-node Distributed System:
Implementation of Berkeley’s algorithm for logical clock synchronization and Vector Clock algorithm for totally ordered multicasting. Project was built in Java.
Bd-Bay System Design:
System design for an online payment-on-delivery system using Use-Case diagram, Class diagram, Collaboration diagram, Sequence diagram and State-chart diagram.
Publications
NuSegDA: Domain Adaptation for Nuclei Segmentation
Mohammad Minhazul Haq, Hehuan Ma, Junzhou Huang
Frontiers in Big Data - Medicine and Public Health, Volume 6, February 2023
Self-Supervised Pre-Training for Nuclei Segmentation
Mohammad Minhazul Haq, Junzhou Huang
Medical Image Computing & Computer Assisted Intervention (MICCAI), 2022
Adversarial Domain Adaptation for Cell Segmentation
Mohammad Minhazul Haq, Junzhou Huang
Medical Imaging with Deep Learning (MIDL), 2020
Graph Attention Multi-instance Learning for Accurate Colorectal Cancer Staging
Ashwin Raju, Jiawen Yao, Mohammad Minhazul Haq, Jitendra Jonnagaddala, Junzhou Huang
Medical Image Computing & Computer Assisted Intervention (MICCAI), 2020
Extra-curricular Activities
Reviewer for International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) 2022
Reviewer for International Conference on Pattern Recognition (ICPR) 2022
Problem Solver at Leetcode and UVa Online Judge
Participant, Kaggle Data Science Bowl 2017
Volunteer & Code Reviewer, Java Fest 2014, Bangladesh
Contestant, Weekly Programming Contest 2009-2014, BUET
