Rudolph Pienaar, PhD
Technical Director
Fetal-Neonatal Neuroimaging and Developmental Science Center
Staff Scientist
Boston Children's Hospital
introduction
Technical Director
- Lead software design efforts at improving computational research
- Software/cloud compute
- Portability of computation and data
- Algorithmic design
- Design center-wide computing infrastructure
- Compute cluster
- Main backend server
- Various VMs
- Backups, etc
overview
- background
- experience
- research and development
- projects
background
PhD Electrical/Electronic Engineering
- Reinforcement Learning in biomechanical control system (self adapting posture balancer)
PostDoc
- MRI processing
- Some pre-processing
- Mostly post-processing
experience
15+ years in medical software R&D
- Image pre-processing
- Pulse Programming on Siemens MRI scanners
- Software for off-line rapid k-space inverse FFT
- Image post-processing
- DICOM data structures
- Mathematical analysis of reconstructed brain surfaces
- Systems architecture
- Pipeline design
- Visualization on the web
- Remote computation using containerization
experience
- 10+ years experience in team leading
- Lead several small, agile software teams
- Project development cycle in 6 month periods
- Web-based visualization
- xtk
- ami
- Remote computation/coordination
- CHRIS
experience
- 5+ years in industry/acadmia collaborative projects
- Hospital scanner QI
- Cloud-based radiology
- Massachusetts Open Cloud
- GE Healthcare/cloud
- Image data to/from GE cloud
- Normative image comparison
- NVidia
- Academic research partner
- Hardware support for in-house cluster
- Academic research partner
experience
- 5+ years in industry/acadmia collaborative projects
- GE
- Learning factory
- Coolidge.io
- IBM
- Partner with PowerAI tools and evaluation
- IBM Watson
- Nimbix Advanced HPC
- RedHat
- Lead developer with cloud-based scheduling and computation on open cloud
- GE
skills
Neurocomputational research
- Somewhat current
- Research into brain surface analysis
- Research into network connectivity
- experience with
- Machine learning
- Reinforcement learning
- Genetic algorithms
- Machine learning
research and development
- Lead Architect for software system development
- medical computational environments
- understanding data needs and constraints
- understanding computational requirements for medical settings
- clinical
- research
- Design systems to interface with home-grown AI pipelines
- Design systems to interface with industry standard
AI tools and pipelines
research and development
- Design systems to interface with home-grown AI pipelines:
- Allow for rapid deployment of computational tools
- Standardized development process
- Web-based, cloud-based
- All data handled by system
- Powerful visualization libraries for input and output data
research and development
- Design systems to interface with home-grown AI pipelines:
- Predictive algorithms
- NIRS related ML and data mining with Drs Sutin/Lin
- Monte Carlo simulations of light propagation through brain tissue leveraging GPUs
- Prostate nerve image segmentation with University of Magdeburg using U-NN
research and development
- Design systems to interface with industry standard tools:
- Python-based tool chain and environment
- Use available and established data-science frameworks:
- tensorflow
- keras
- opencv-python
- pandas
- scikit-image
- etc
research and development
- GE Healthcare:
- Visualization for Radiology Workbench
- Interface with GE cloud for data storage
- Interface with GE AI tools:
- Learning Factory
- Route 66
- Coolidge
On going and deepening collaboration for systems development.
research and development
- Nvidia:
- Tensor-flow libraries
- Nvidia openlabs
Use of lower-level tools and libraries for specific neural-network training in image analysis tasks.
research and development
- IBM:
- PowerAI systems and software
- Collaboration on the Massachusetts Open Cloud
- Collaboration on the Massachusetts Open Cloud
- PowerAI systems and software
Use IBM available AI tools for object / scene recognition and develop framework to connect to medical data.
research and development
- RedHat, Inc:
- Collaboration on the Massachusetts Open Cloud
- Openstack / kubernetes
- Openswift
Develop backend opensource infrastructure to use framework on industry standard cloud systems.
medical compute... in the cloud
definite need for data mobility and centralization...
with compute being "free" to process data where-ever it might naturally collect...
CHRIS...
CHRIS is a bio-medical data workflow manage that allows easy and intuitive collection, analysis, and sharing of data between parties.
- allows for advanced view rendering
- allows for sharing of images
- allows for immersive real time collaboration
- allows for post-processing of images
many data sources... many compute sources...
via here
connect data here
to here...
and compute!
overview / goals
how do I run this on the cloud?
how do I run this on "someone else's computer"?
system topology
Main web server local data repository
remote compute nodes
data source (e.g. in hospital PACS)
UI concepts
UI concepts
UI concepts
Rav...
Custom web-based viewer...
BU partnership
Boston University Computer Science
Mentoring students on Cloud computing
Academic partner on the MOC
RedHat partnership
RedHat is active partner developing CHRIS infrastructure to the MOC
CHRIS as platform to help foster community and opensource development in the cloud using this paradigm
RedHat interested in helping use CHRIS to be a platform for "hack-a-thons" based on openswift, kubernetes, etc.
MOC
references... ChRIS
-
https://github.com/FNNDSC/ChRIS_ultron_backEnd
-
https://github.com/FNNDSC/ChRIS_ultron_frontEnd
references... services
-
https://github.com/FNNDSC/viewerjs
-
https://github.com/FNNDSC/gcjs
-
https://github.com/FNNDSC/fmjs
-
https://github.com/FNNDSC/toolbarjs
-
https://github.com/FNNDSC/rboxjs
-
https://github.com/FNNDSC/rendererjs
-
https://github.com/FNNDSC/thbarjs
-
https://github.com/FNNDSC/pman
-
https://github.com/FNNDSC/pfioh
-
https://github.com/FNNDSC/pfcon
references... viewers
-
https://goxtk.com
-
https://github.com/FNNDSC/ami
team
- Jorge Bernal
- Nicolas Rannou
- Rudolph Pienaar
- Yangming Ou
- Daniel Haehn
- Daniel Ginsburg
- Ellen Grant
BCH Team
BU Team
- Orran Krieger
- Ata Turk
- Aditya Awalker
RedHat
- Dan McPherson
- and others!
fin!
Thank you!
self_overview_2
By Rudolph Pienaar
self_overview_2
Quick intro / overview
- 902