The classes have been posted on YouTube on a dedicated playlist.
For Day 2 - Part 3 (Connecting with IceGadgetron), please visit the Siemens IDEA board for the link.
Some email addresses cannot be reached (...@gmial.cpm), we try to correct them, but some of them are more difficult to guess. For other, it might also be a security rule from your mail server. If you haven't received the instruction by email and most importantly the Slack invitation within a few day. Please contact us gadgetron2020 at sciencesconf dot org
During June 2020 we will host the “GadgetronOnlineClass”. Although initially scheduled to take place in Bordeaux as a Summer School, the COVID-19 situation made that impossible. We are instead hosting it online.
This course is aimed at both new and experienced users of Gadgetron, covering basic reconstruction as well as the latest functionalities. The topics covered are intended for researchers in basic science and/or clinical research. The online course will also provide examples of how to use the vendor-agnostic MRD file format (formerly “ISMRMRD”) for your custom reconstructions.
The Gadgetron Online Course is made of several modules that constitute together the scientific stack. Gadgetron made use of standard computing languages (C++, Cmake, CUDA, Python, Matlab) that are themselve calling hundreds of libraries/packages/functions. We won't cover everything in this short course, but you should get enough information to decide if your research can benefit from Gadgetron. And I bet it will likely do.
For any questions, feel to contact us on the forum of Gagdetron or at gadgetron2020 /at/ sciencesconf.org
Organizers: David Hansen, Kristoffer Knudsen, Hui Xue, Oliver Josephs, Vinai Roopchansingh, John Derbyshire, Adrienne Campbell, Rajiv Ramasawmy, Aurélien Trotier, Stanislas Rapacchi, Maxime Yon, Pierre Bour, Valéry Ozenne
Please follow the registration link. Registration is free. We will use the platform dedicated to the initial summer school for the registration process and email communication. Otherwise please refer to the GitHub website GadgetronOnlineClass. Thanks for your comprehension.
Note that the time are defined in the Central European Summer Time (CEST) zone.
Date | Time | Link | Topic | Tutor |
---|---|---|---|---|
June 11, 2020 | 13:45-14:00 | [Welcome] | Valéry Ozenne | |
June 11, 2020 | 14:00-14:30 | link | [A tour of Gadgetron] | David Hansen |
June 11, 2020 | 14:30-15:30 | link | [Practical introduction to Gadgetron] | Kristoffer Knudsen |
June 11, 2020 | 15:30-15:40 | [Break] | ||
June 11, 2020 | 15:40-17:00 | link | [Hand's on session: Basic reconstruction using Python] | Valery Ozenne |
June 11, 2020 | 17:00-18:00 | link | [Q&A] | Everybody |
Date | Time | Place | Topic | Tutor |
---|---|---|---|---|
June 18, 2020 | 14:00-15:00 | link | [MRD Part 1: Introduction] | Maxime Yon |
June 18, 2020 | 15:00-15:55 | link | [MRD Part 2: ]Siemens/GE/Bruker raw data conversion to MRD through XML style sheets, and working with HDF5 files] | Vinai Roopchansingh & J. Andrew Derbyshire |
June 18, 2020 | 15:55-16:05 | [Break] | ||
June 18, 2020 | 16:05-17:00 | link | [Communication process with the Siemens scanner] | Hui Xue |
June 18, 2020 | 17:00-18:00 | link | [Q&A] | Everybody |
Date | Time | Place | Topic | Tutor |
---|---|---|---|---|
June 25, 2020 | 14:00-15:00 | link | [External-Language-Interface ] | Kristoffer Knudsen |
June 25, 2020 | 15:00-15:55 | link | [Protyping at the scanner with MATLAB part 1] | Aurélien Trotier & Stanislas Rapacchi |
June 25, 2020 | 15:55-16:05 | [Break] | ||
June 25, 2020 | 16:05-17:05 | link | [Protyping at the scanner with MATLAB part 2] | Oliver Josephs |
June 25, 2020 | 17:05-18:00 | link | [Q&A] | Everybody |
Date | Time | Place | Topic | Tutor |
---|---|---|---|---|
July 2, 2020 | 14:00-15:00 | link | [How to write a C++ Gadget ] | David Hansen |
July 2, 2020 | 15:00-15:55 | link | [The Generic Cartesian Chain and toolboxes] | Hui Xue |
June 2, 2020 | 15:55-16:05 | [Break] | ||
July 2, 2020 | 16:05-17:00 | link | [How to integrate your AI model inline on the scanner] | Hui Xue |
July 2, 2020 | 17:00-18:00 | link | [Q&A] | Everybody |
To achieve good interaction between lecturers and participants - and especially tutors and participants in the work in progress sessions - we highly recommand to read the preparation list.
Preparation list for speakers is available here.
Preparation list for participants here.
All information will be written in a README file in each directories. Additionnal contents like data, codes, powerpoint will be uploaded or indicated in the material folder for each lecture
├── Courses
│ ├── Day1
│ │ ├── Lecture1
│ │ │ └── README.md
│ │ ├── Lecture2
│ │ │ └── README.md
│ │ └── Lecture3
│ │ └── README.md
│ ├── Day2
│ │ ├── Lecture2
│ │ │ └── README.md
│ └── Day3
├── Installation
│ └── README.md
├── Interactive-Sessions
│ └── README.md
├── Preparation
│ └── README.md
└── README.md
Since all participants are working at home on their own computer, we asked the participants to test their MRD and Gadgetron installation in advance. Detailed installation instructions has been summarized here.
Feel free to contact us and to post any inquiries on the gadegtron forum
Note that this course is based on the following teaching material: Tutorial with docker [link](http://gadgetron.github.io/tutorial/) Tutorial Hello word [link](https://github.com/gadgetron/gadgetron/wiki/Gadgetron-Hello-World) Running one of them before is highly recommended
To follow along on your own computers with the live demonstrations and coding sessions, the data for most of those can be found here
If you prefer to search the dataset by its DOI on Zenodo, it is: 10.5281/zenodo.3888658
The final link these data can be accessed at, which should bring you back to the same dataset, is https://doi.org/10.5281/zenodo.3888658
This file should have a checksum of: 361e2dccec2949c4914414dada6d2b5dd841aff4 computed using 'shasum' on macOS and openSUSE.
A second archive file for additional data for the class has been added here.
The DOI link for these data is https://doi.org/10.5281/zenodo.3906695.
The checksum computed by shasum for this second archive is: ddfd703098bde34f95dc886aa268d0c59d5de5da
This lecture does not attempt to be comprehensive and cover every single feature, or even every commonly used feature. Instead, it introduces many of Gadgetron's most noteworthy features, and will give you a good idea of Gadgetron's capability and usage.
See also:
This lecture is a Practical Introduction to Gadgetron. It takes a very hand-on approach to getting started with Gadgetron, aimed at giving new users the information they need to assemble and run their own reconstructions using Gadgetron.
This lecture will cover starting and running Gadgetron, controlling Gadgetron behaviour through configuration files, and provide a sensible introduction to a handful of very common Gadgets.
See also:
The primary goal of this lecture introduces the python gadget and the ismsmrd-python-toolboxes that contains various toolboxes dedicated to common issues. Its different submodules correspond to different applications, such as fourier transfrom, grappa reconstruction, etc.
2. MRD, a vendor-agnostic file format for MRI reconstruction (or toward the scanner room) (day 2: June 18th)
This lesson introduces the ISMSMRD format
See also:
2.2 - Siemens/GE/Bruker raw data conversion to MRD through XML style sheets, and working with HDF5 files
This session discusses how the conversion process in ISMRMRD utilizes style sheets to convert the vendors' proprietary formats and data values to the fully and openly documented ISMRMRD format. There will also be an interactive demonstration of working with HDF5 files in a Jupyter notebook environment with Python 3.
See also:
This course servers an introduction to the IceGadgetron software. The main functinality of IceGadgetron is to stream siemens MR kspace data to MRD formant and send data to Gadgetron. As Gadgetron reconstructed images and send images back to scanner, IceGadgetron will receive the images and pass them down the functor chain. In this way, the inline processing is achieved, which allows seamless integration of Gadgetron to MR scanner and be used by non-technical users.
See also:
This lesson introduces connecting gadgetron to matlab.
See also:
This lesson introduces ...
See also:
This lesson introduces ...
See also:
This lesson introduces ...
See also:
This lesson introduces the generic reconstruction concept and implemenation in Gadgetron. The generic recon is first an idea to develop a systematic handling of multi-dimension kspace MR data and implemented most commonly used technique to support multiple imaging protocols. The key idea is the "on-the-fly" triggering and reconstruction.
The key components include data triggering, data bucket organization, auto-calibration data preparation, coil compression, reconstruction, kspace filter, partial fourier handling, and FOV adjustment gadgets. User can specify which dimension to trigger (e.g. for a multi-slice cine scan, triggering can be done with SLICE dimension). In this way, the reconstruction can start after data from one slice is acquired, instead of waiting to the end of scan. After triggering, every data bucket was processed in a uniformed manner and each encoding space is handled separately.
The corresponding gadgets also server as good starting point for user to overload for their own use cases, if the generic chain may not handle a specific use case.
This lesson introduces the topic of neural network model in Gadgetron.
The introduction of AI model into Gadgetron is a recent development, but of great importance to improve imaging, computing, analysis and reporting. First, Gadgetron offers ability for user to load their pre-trained AI models and apply models to incoming kspace and imaging data. Second usecase of Gadgetron to support AI is to allow user to do inline training and apply resulting model to incoming data (e.g. check the Grappa_AI example, where a Pytorch implemention of Grappa reconstruction was demonstrated).
User can choose to implement AI functionalities in python gadgets or in python functions which can be called in C++ gadgets.
See also:
The driving themes of the Gadgetron are notably cardiac imaging, interventional imaging, MR-PET imaging, high field and low field imaging. Here are some works done using the Gadgetron. The list is non-exhaustive.
- kt-SENSE, non-Cartesian, iterative SENSE, sequence: golden angle radial bSSFP projections. link
- 3D l1-SPIRiT Reconstruction on Gadgetron based Cloud. link
- Integration of the BART into Gadgetron for Inline Cloud-Based Reconstruction. link
- MR Fingerprinting using a Gadgetron-based reconstruction . link
- Cardiac MR Fingerprinting in Gadgetron for Online Reconstruction. link
- TPVM Tissue Phase velocity mapping, Spiral trajectory. link
- Spiral imaging with off-resonance reconstruction. link
- 4D DCE MRI, Free-Breathing Liver Perfusion. Spiral trajectory + GRAPPA. link
- SMS T1 mapping. link
- Multi-vendor Hyperpolarised 13C analysis . link
- Full Free-Breathing Protocol for CMR. link
- In-line cardiac perfusion mapping link, link, link
- In-line cardiac perfusion and deep learning link
- MR-PET imaging : SIRF link
- MR-PET imaging : respiratory and cardiac motion correction, reco: joint Compressed Sensing reconstruction. link
- MR-PET imaging : 4D CBCT-based proton dose calculation. link
- Temperature mapping. link, link
- ARFI: sequence single-shot gradient EPI, reco GRAPPA. link
- Highly accelerated cardiac cine : bFISTA‐SPIRiT et bFISTA‐SENSE. link
- Low field magnetic resonance imaging scanner for cardiac imaging. link
- Augmented Reality link, link
- Cardiac diffusion, GRAPPA. link
- Neuro Diffusion,3D multi-shot-EPI avec PF. link
- Coronary magnetic resonance angiography, variable density sampling + Compress Sensing. link
- Real-time feedback 3DEPI fMRI, Matlab SENSE reconstruction link