Skip to content

raidionics/Raidionics-models

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

Models and resources for use in Raidionics and Raidionics-Slicer

All models are in ONNX format, compatible with the latest Raidionics version (v1.2) and Raidionics backend (V1.1.0).

1. MRI models

  • Preoperative glioblastoma, diffuse low-grade glioma, meningioma, and metastasis segmentation: the models have been described in Preoperative Brain Tumor Imaging: Models and Software for Segmentation and Standardized Reporting

  • "_multiclass" models, trained with the AGU-Net architecture over the BraTS challenges' datasets, and providing three output labels: contrast-enhancing tumor, necrosis, and edema.

  • Early postoperative glioblastoma segmentation: multiple models leveraging different sets of full-volume inputs, introduced in Segmentation of glioblastomas in early post-operative multi-modal MRI with deep neural networks.

    • MRI_GBM_Postop_FV_1p model: Uses one MR sequence as input: postop T1-CE.
    • MRI_GBM_Postop_FV_2p model: Uses two MR sequences as inputs: postop T1-CE, postop T1-w.
    • MRI_GBM_Postop_FV_3p model: Uses three MR sequences as inputs: postop T1-CE, postop T1-w, postop FLAIR.
    • MRI_GBM_Postop_FV_4p model: Uses three MR sequences and a prediction as inputs: postop T1-CE, postop T1-w, preop T1-CE, preop tumor mask.
    • MRI_GBM_Postop_FV_5p model: Uses four MR sequences and a predictions as inputs: postop T1-CE, postop T1-w, postop FLAIR, preop T1-CE, preop tumor mask.
  • Sequence classification: not validated nor introduced in any previous publication yet. Classifies between T1-weighted (T1w), gadolinium-enhanced T1-weighted (T1w-CE), T2-weighted fluid attenuated inversion recovery (FLAIR), and T2-weighted (T2).

2. CT models