Hi, I'm Hans J. Johnson (@hjmjohnson) hans-johnson@uiowa.edu
🎓 Professor Electrical & Computer Engineering — University of Iowa
🚀 *Software Engineer • Open Science Advocate • Research Problem Solver *
💡 Medical Image Processing • HPC • Software Engineering for Clinical AI
I am a professor and research software engineer whose passion lies in accelerating scientific discovery through reproducible, scalable, and well-engineered computational methods.
My work bridges the worlds of biomedical imaging, machine learning, software engineering, and high-performance computing (HPC) to create sustainable software infrastructures for translational research.
I'm a passionate software developer dedicated to creating innovative solutions and contributing to open-source projects. I love tackling complex problems and building tools that make a difference.
- 🔭 I'm currently working on cutting-edge software projects
- 🌱 I'm continuously learning new technologies and best practices
- 👯 I'm looking to collaborate on impactful open-source initiatives
- 💬 Ask me about software development, algorithms, and system design
My research aims to accelerate discovery through efficient analysis of large-scale, heterogeneous, multi-site biomedical datasets.
To achieve this, I lead efforts that combine:
- ⚙️ Software Engineering — reproducible, version-controlled pipelines
- 🧠 Machine Learning — AI-assisted medical image segmentation and prediction
- 💾 High-Performance Computing (HPC) — massive parallelization and distributed analysis
- 🌐 Informatics — secure and automated multi-site data harmonization
I direct interdisciplinary projects that transform modern HPC resources into collaborative discovery platforms spanning many-core systems, distributed storage, and centralized repositories.
| Role | Project | Description |
|---|---|---|
| 🧩 President | Insight Software Consortium | Oversees ITK — the leading open-source toolkit for medical image analysis, advancing software education and sustainability. |
| 🧬 Software Engineering Lead | NA-MIC | Contributing to one of eight NIH National Centers for Biomedical Computing to promote open, interoperable imaging tools. |
| 🧠 Executive Committee Member | PREDICT-HD | 28-site, 11-year longitudinal Huntington’s Disease study tracking 1200+ subjects — leading informatics and imaging infrastructure. |
| 🧪 Co-PI | NeuroNEXT | Data coordinating center for a 27-site NIH clinical trial network — developing multi-modal data capture and QC systems. |
| 🌍 Steering Committee & Imaging Lead | TRACK-HD | International 4-site imaging study — architected automated data collection and harmonization pipelines. |
The future of biomedical discovery depends on our ability to efficiently leverage HPC and informatics across disciplines.
My work demonstrates how rigorous software engineering practices and HPC infrastructures enable reproducible analysis at scale.
- Managed 32 international sites collecting imaging, genetic, and behavioral data from 1200+ participants annually.
- Designed a centralized electronic data capture and protocol management system integrated with XNAT.
- Created custom QC tools for efficient visual validation and feedback loops to data collection sites.
- Established infrastructure that remains in regular use over a decade later.
To analyze the massive PREDICT-HD dataset (1346 subjects, 3713 sessions, 10,000+ images), we:
- Prototyped automated longitudinal pipelines using ITK, SimpleITK, and custom Python/C++ tooling.
- Leveraged University of Iowa HPC clusters (Helium/Neon/Argon) for distributed computation.
- Consumed >15,000 CPU-days per month, equivalent to ~$49K/month in commercial compute costs — enabling otherwise infeasible analysis.
These efforts led to:
- Publications in Lancet Neurology and related journals on HD progression modeling.
- New machine learning-based segmentation pipelines for brain MRI, extending to stroke imaging and prognosis prediction.
- Foundational infrastructure for subsequent clinical AI pipelines in NIH-funded research.
- Migration of legacy imaging codebases (e.g., Qt5 → Qt6, SWIG → pybind11)
- CI/CD and automated test coverage for C++/Python hybrid systems
- Cross-platform packaging (macOS ARM64, Linux HPC, Windows MSVC)
- FAIR data pipelines for clinical imaging research
- Automated harmonization of multi-modal datasets via ML-driven inference
| Domain | Tools & Technologies |
|---|---|
| Programming | C++, Python, Bash, Java, CMake |
| Frameworks | Qt6, PyTorch, MONAI, ITK, SimpleITK |
| DevOps & CI/CD | GitHub Actions, GitLab CI, Docker, act, clang-tidy |
| Visualization | matplotlib, napari, ParaView, 3D Slicer |
| Compute Environments | Slurm, NFS, VMware, macOS ARM, Linux HPC |
| Version Control | Git, Git LFS, pre-commit hooks, semantic versioning |
| Repository | Focus | Highlights |
|---|---|---|
| ITK | Core C++ image analysis library | 1st of 83 top contributors (ITK v4), CI modernization |
| SimpleITK | Simplified ITK for rapid prototyping | Multi-language bindings via SWIG |
| BRAINSTools | Neuroimaging pipeline suite | Atlas building, DICOM I/O, quantitative MRI |
| PythonQt | Qt ↔ Python bridge | Migration to Qt6 + CMake modern build system |
| NeuroPredAI | Clinical AI model development | MONAI-based GPU training for brain segmentation |
My publication record reflects a deep integration of software engineering, imaging science, and AI for clinical impact.
- PREDICT-HD Consortium — longitudinal analysis of neurodegeneration (Lancet Neurology, NeuroImage, Biological Psychiatry)
- TRACK-HD / TRACK-ON-HD — automated pipelines for progression modeling
- ITK / SimpleITK — reproducible frameworks for open science imaging
- Recent Themes: Machine learning-based segmentation, harmonization of multi-site MRI data, and AI-assisted clinical decision support.
💡 I’m always open to collaborations that advance scientific software sustainability and reproducibility.
📫 Contact
