@jerome_esteves@raimcsantos@3DSlicerApp Here are the 3D image files of a liver CT and the fully automatic segmentation results (obtained in 2 minutes): https://t.co/T9tVcMd1Mk. Source CT image: Medical Decathlon data set - Task03_Liver - imagesTr/liver_100.nii.gz
@jerome_esteves@raimcsantos@3DSlicerApp Not a lot of time and pain to save here. You can get this segmentation in 3D Slicer with TotalSegmentator extension within 2 minutes, fully automatically. It is completely free and comes without any licensing restrictions. The data does not leave your computer. No GPU required.
A small new feature in @3DSlicerApp for painting segmentations into 3D images at higher quality. Developed for #SlicerHeart but has many other applications. More info: https://t.co/wtAoFSweCh
@ayomi4_adebayo@3DSlicerApp In general, you run the server in a separate Python environment outside Slicer. After you started the server, Slicer can connect to it.
@ayomi4_adebayo@3DSlicerApp NVIDIA-AIAA developers stopped working on the project and moved on to develop MONAILabel, so we gave up on maintaining a public NVIDIA-AIAA server.
@MarkLikesTrauma TotalSegmentator is trained on a wide variety of data, so that should work on pediatric cases, too. There are several MONAILabel segmentation tutorials and examples and you can ask the @ProjectMONAI community if you need help with any specific questions.
@medslicer@3DSlicerApp You don't need to develop any new code but you have to find and follow tutorials and documentation to run the training scripts. You also need powerful GPU for training.
This is a game changer. Segment 100+ structures in any whole-body CT image in 2 minutes using TotalSegmentator in @3dslicerapp
All free, open-source software. Runs on any computer, no GPU is required. See more information at https://t.co/oH1vjz3KoJ
@medslicer@3DSlicerApp You can submit corrections if you find inaccuracies in the original training data set. For other images, you probably need to retrain the network yourself.
@basalmind@alexandrecadrin Once you have segmented the vasculature, you can separate the arterial and venous branches by placing a few seed points and growing from those seeds, inside the vessels (using in Segment Editor in 3D Slicer - https://t.co/MoGhTLmc2G)
@MikolajBuchwald@szym0nk We have specialized presets for MV, TV, CAVC, LAVV, and a generic valve preset. You can use the generic valve or MV preset for aortic valve. You could also create a new aortic valve quantification preset by specifying reference points and measurements based on them.
We have released a number of tools for cardiac valve modeling, quantification, simulation and we have many more in the pipeline. All free, freely usable, open-source software.
4/5) The multimodality-capable (3DE, CT, CMR) tricuspid valve modeling tools we created for this study are now available open-source @3DSlicerApp#SlicerHeart#SlicerSALT
@aigonewrong@3DSlicerApp It would be also nice to see Hausdorff distance results. Dice is widely used (probably because it is easy to compute) but works very poorly for thin, small, or elongated segments.
@aigonewrong@3DSlicerApp Interesting results, thanks for sharing. Pediatric CTs were not used for TotalSegmentator training, so some inaccuracy is expected. You could consider adding them to the TotalSegmentator training data set and retrain to see if it can handle pediatric cases, too.