This training school is online, with presentations via Teams. Participants are asked to join the SIRF Discord server (registration link, channel for the training school) for interacting with the instructors and other participants.
This school will start with introductory presentations, followed by a sequence of notebooks of the SIRF-Exercises and CIL-Demos, and final presentations.
For each notebook, there will be a brief introduction, 30 minutes where participants are expected to go through the notebook (ideally in groups), finalising with a brief recap for everyone.
10:00- 10:30 Intro presentation
- What is SIRF?
- Basics of PET
10:30 – 11:00
Walk through SIRF Introductory/introduction notebook to show how to create image data objects for MR, CT and PET and how to work with them.
11:00 – 11:40
SIRF notebook PET/image_creation_and_simulation shows how to use SIRF to create images via geometric shapes and forward project them with and without attenuation. Exercises extending the simulation to include noise and other parts of the PET model are suggested.
11:40 – 12:20
SIRF notebook PET/OSEM_reconstruction shows how to implement Ordered Subsets Expectation Maximization (OSEM) in SIRF and suggests some exercises for reconstruction with and without attenuation etc.
Lunch break
13:00-13:40
SIRF notebook PET/ML_reconstruction shows how to monitor progress of a SIRF reconstructor (currently using OSEM as an example) and implement a (simplistic) gradient-ascent algorithm using SIRF. This notebook can be extended to use regularized reconstruction as well.
13:40-14:20
SIRF notebook PET/DIY_OSEM invites you to implement Maximum Likelihood Expectation Maximization (MLEM) and OSEM in SIRF yourself using SIRF functionality.
14:20-15:00
CIL notebook PyData22_deblurring gives a brief introduction in optimisation in inverse problems with CIL with an example of image deblurring.
15:00 – 15:30
What is PETRIC2 and how to participate? (Kris)
15:30 – 16:00
Feedback and closing (Edo)