Brainchop.org is a browser-based 3D MRI rendering and segmentation tool that simplifies volumetric analysis of structural MRI. It processes full brain volumes in one pass using the MeshNet segmentation model, offering extensibility via an open-source framework. Brainchop performs client-side processing, ensuring privacy, and utilizes hardware acceleration across different brands and architectures.
Version
3.0.0
Date
March 28, 2024
Description of breaking changes
Brainchop can now run powerful models on weaker devices. The Brainchop Models Zoo has three classes of models, as shown in the attached gif:
Flash filet is a small, fast model that works well on data similar to HCP structural MRIs. You can find it in the menu as "Tissue GWM (light)" with a lightning bolt symbol (⚡).
Thin slice models are high-quality but fragile if input T1s are highly variable. They are temporarily marked with a 🔪 icon in the menu to indicate their thin slice class.
Rough Chop is a new model for brainchop.org. It produces reasonable results on various data types, including clinical scans and infant data. 🪓
Open Source Project name
Brainchop
What is your project?
Brainchop.org is a browser-based 3D MRI rendering and segmentation tool that simplifies volumetric analysis of structural MRI. It processes full brain volumes in one pass using the MeshNet segmentation model, offering extensibility via an open-source framework. Brainchop performs client-side processing, ensuring privacy, and utilizes hardware acceleration across different brands and architectures.
Version
3.0.0
Date
March 28, 2024
Description of breaking changes
Brainchop can now run powerful models on weaker devices. The Brainchop Models Zoo has three classes of models, as shown in the attached gif:
Flash filet is a small, fast model that works well on data similar to HCP structural MRIs. You can find it in the menu as "Tissue GWM (light)" with a lightning bolt symbol (⚡).
Thin slice models are high-quality but fragile if input T1s are highly variable. They are temporarily marked with a 🔪 icon in the menu to indicate their thin slice class.
Rough Chop is a new model for brainchop.org. It produces reasonable results on various data types, including clinical scans and infant data. 🪓
GitHub Repo
https://github.com/neuroneural/brainchop
Website
brainchop.org
Link to changelog
https://github.com/neuroneural/brainchop/releases/tag/v3.0.0
Social media
https://twitter.com/PlisSergey/status/1773349780736876554
Anything else to add?