I’ve been working on a major update to SlicerModalityConverter.
Here’s a video introducing the new features, with two use cases: generating synthetic CT from a T1-weighted brain MRI and synthetic PET from a chest CT scan.
This release focuses on improving the inference workflow in 3D Slicer, with several practical additions:
• An improved inference process that speeds up local inference while keeping 3D Slicer responsive.
• Support for remote inference, including HTTP/HTTPS connections and authentication.
• Real-time monitoring of local and remote resources, with the ability to select resources on the remote machine.
If you would like to integrate your own model for personal use on your local machine, or make it available to the 3D Slicer community through the extension, the GitHub repository includes a guide to help you get started.
The goal is to make medical image synthesis more accessible to researchers and developers working with AI and medical imaging in 3D Slicer.
I hope you’ll find it useful. Feedback and contributions are always welcome! 