Curriculum Vitae
Snawar Hussain
PhD researcher working at the intersection of AI, computational imaging, numerical simulation, and physics. I build physics-aware and data-driven methods for modeling complex systems, with broader interests in scientific reinforcement learning, simulation, generative modeling, and computer vision.
Selected Projects
Selected research and engineering projects across computational Physics, Simulation AI-assisted imaging, reconstruction, computer vision, representation learning, and biomedical image analysis.
Hybrid MRI Simulator for Physics-Based Image Formation
Built a hybrid MRI simulator that combines efficient phase-graph propagation with Bloch-resolved slice-selective RF operators. The framework simulates sequence-driven signal evolution, relaxation, gradient encoding, and image formation, providing a physics-based foundation for fast MRI experiments and AI-assisted sequence optimization.
Facial Landmark and Depth Estimation Inside the MRI Scanner
A computer vision project for estimating facial landmarks and depth information inside the constrained MRI scanner environment. The system focuses on robust face-centered tracking under partial occlusion, restricted viewpoints, and scanner-specific imaging constraints, enabling motion-aware analysis during MRI acquisition.
gSTARparser: Scanner-Runnable MRI Sequences for Simulation
A project showcase for interpreting scanner-runnable gammaSTAR sequence programs into simulator-ready blocks. The parser preserves RF events, gradient timing, ADC placement, and execution structure, enabling direct use of vendor-agnostic sequence descriptions inside MRI simulation workflows.
KIMBI: AI-Supported Intelligent Magnetic Resonance Imaging
Contributed to the KIMBI project in the context of Fraunhofer MEVIS and the U Bremen Research Alliance. My work focused on MRI simulation components compatible with gammaSTAR, a vendor-agnostic MR sequence development framework, supporting the connection between executable sequence descriptions, simulation, and AI-assisted imaging workflows.
ROBUST: Rodent Observable Behavior Understanding and Study Toolkit
A toolkit for unsupervised behavior analysis of rats using self-supervised representation learning with variational autoencoders and Hidden Markov Model based behavioral state clustering.
2D/3D Pose Estimation
Developed a multi-view pose estimation workflow that reconstructs 3D pose from multiple 2D camera views using geometric triangulation and camera-view correspondence.
Retinal Blood Vessel Segmentation
Implemented a U-Net based segmentation pipeline for retinal blood vessel extraction, including model training, evaluation, and qualitative visualization of segmentation results.
Education & Research Experience
PhD Candidate in Physics
Fraunhofer MEVIS / University of Bremen
Working on physics-based simulation, AI-assisted computational imaging, MRI sequence modeling, and efficient numerical methods for scientific computing.
Research Scholar in Computer Science
Zhejiang University
Research experience in machine learning, computer vision, representation learning.
M.Sc. in Control Science and Engineering
Central South University
Focused on intelligent systems, pattern recognition, machine learning, and computational methods for biomedical imaging and image processing.
B.Sc. in Mechatronics Engineering
National University of Sciences and Technology
Foundation in robotics, embedded systems, control, electronics, and engineering computation.
Publications
My research contributions span across simulation, computational imaging, scientific machine learning, and computer vision.
View Google Scholar