Academic & Clinical Research
Research & Clinical AI Work
Investigating reward-driven neural plasticity algorithms for medical image segmentation and Attention U-Net architectures.
Graduate Researcher - Master's Thesis
Graduate Researcher | Bahir Dar UniversityThesis Title: Reward-Driven Neural Plasticity Inspired Optimization for Enhancing U-net Based Medical Image Segmentation
- Designed a biologically inspired optimization algorithm based on neural plasticity and reward-driven learning.
- Developed deep learning pipelines using PyTorch and TensorFlow for brain MRI segmentation (tumor & stroke lesions).
- Implemented preprocessing workflows including normalization, resampling, and augmentation for NIfTI and DICOM datasets.
- Evaluated performance using Dice Coefficient, IoU, Precision, and Recall metrics.
- Achieved improved convergence stability and segmentation accuracy compared to Random Search and Genetic Algorithms.
Neuro-Inspired U-Net Optimization Project (Open Source)
Lead Developer | GitHub Project- Developed a reproducible Python framework integrating bio-inspired optimization into deep learning training.
- Implemented Attention U-Net models for fine-grained medical image segmentation.
- Built automated medical imaging pipelines using MONAI and SimpleITK (skull stripping, preprocessing, artifact removal).
- Designed modular codebase for research reproducibility and extension.
Interactive Lab DemoMONAI & PyTorch Pipeline
Brain MRI Deep Learning Segmentation Visualizer
Simulated high-resolution axial T1/T2 Brain MRI segmentation using bio-inspired plasticity neural optimization.
Dice Similarity (DSC)0.942
Jaccard Index (IoU)0.891
Precision0.954
Inference Latency12ms
MRI Slice DepthZ-Axis: 72 mm
Segmentation Mask Opacity75%
● Tumor Core● Edema● Ventricle
Primary Research Focus
Medical Image Analysis (MRI / CT segmentation, neuroimaging)
Bio-inspired and Neuro-inspired Learning Systems
Explainable AI (XAI) for Healthcare Systems
Deep Learning Optimization and Architecture Design
Domain Adaptation in Clinical AI
Pediatric Neuroimaging and Tumor Analysis