Sepand AliMadadSoltani

Education

• Concentration: Medical Image Processing
• GPA: 15.31/20
• Courses: Magnetic Resonance Imaging (MRI), Segmentation & Registration, Artificial Intelligence in Medical Imaging, Image Reconstruction & Inverse Problems
  
• €10,000 Excellence Scholarship awarded for excellent academic background
  

• Concentration: Biomedical Engineering
• GPA: 16.33/20
• Courses: Statistical Pattern Recognition, Signals & Systems, C Programming, Engineering Math, Engineering Probability

Research Experience

• Developed non-invasive alternative to arterial sampling for PET quantification in 59 patients
• Built an accurate and automatic carotid artery segmentation pipeline on TOF-MR angiography images
• Implemented Bayesian inference model using Markov Chain Monte Carlo (MCMC) for Partial Volume Correction
• Reduced quantification bias significantly compared to standard image-derived methods (MAPE 13% vs 24%)
• Validated method using Monte Carlo PET simulations and cross-site collaboration with Paris-Saclay

• Built medical image analysis and visualization software in Python
• Implemented intelligent scissors algorithm for semi-automatic boundary detection on 2D image slices
• Integrated the algorithm into the app for fast tissue segmentation with minimal user input
• Incorporated complementary segmentation tools (polygon selection, thresholding, otsu filter)

Publications

Skills

• Programming: Python, C, C++, CMake, Bash
• Software and Tools: GNU/Linux, Git, FMRIB FSL, SPM, 3D Slicer, TPCCLIB
• Libraries: Tensorflow, PyTorch, NumPy, pandas, scikit-learn, Matplotlib, ITK, VTK, Qt
• Languages: English (fluent, TOEFL score:101/120), French (Intermediate), Persian (Native)

Work Experience

• Designed and developed a modern interface using the Qt Framework's QML language
• Built and optimized backend logic in C++ to handle large volumes of data efficiently

• Engineered a node-based workflow for composition of custom AI/ML medical image analysis pipeline in Python analysis
• Built a 3D visualization engine using VTK to render volumetric medical data

Projects

• Processed raw fMRI/MRI data from the ADNI database using FSL (motion correction, registration) to extract functional connectivity maps
• Evaluated RCNN and CNN models in Tensorflow for feature extraction to classify Alzheimer’s disease stages
• Identified critical limitations in using standard deep learning architectures for fMRI analysis

• Implemented brain extraction from structural reference MR image
• Implemented fMRI pre-processing including motion correction, slice timing correction, spatial smoothing, and co-registration
• Optimized the pipeline with parallel processing to accelerate computation on the ADNI dataset

References