This tool is based on a nnU-Net model trained and evaluated on more than 700 CT and CBCT scans. It has been shown to be robust in most situations, even in the presence of metallic artefacts or varying field of views.
Segment editor tools and model export features are directly available in the module. The extension does not require a GPU, but a CUDA-capable GPU is needed for fast results (around 1-2 minutes).
Install tutorial and demonstration video:
If you use DentalSegmentator for your work, please cite our paper and nnU-Net:
Dot G, Chaurasia A, Dubois G, et al. DentalSegmentator: robust deep learning-based CBCT image segmentation. Published online March 18, 2024:2024.03.18.24304458. doi:10.1101/2024.03.18.24304458
Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods. 2021;18(2):203-211. doi:10.1038/s41592-020-01008-z
Thank you very much for creating the module, but it doesnât work. I have the latest version of the program installed, as well as all the required plugins. However, when I click âapply,â it just keeps loading and doesnât do anything. Iâve left it running for up to an hour, but nothing happens.
Awesome work. Congrats!
I would like to know if it is possible to set parameters so we can run it on a 2gb NVIDIA laptop GPU.
I run it on CPU just fine (it takes several minutes). But I ran out of memory when tried to run on CUDA.
A GPU with 2GB RAM is only sufficient for some very lightweight visualization tasks, it is too small to be useful for 3D segmentation.
You can either upgrade to a stronger GPU (I would recommend minimum 12GB GPU RAM, but 24GB would be safer if you want to use it for a few years) or use your CPU.