Collection of neural network models for metabolic tumor volume delineation in (Non-Hodgkin) lymphoma patients in FDG-PET/CT images. Intended to use within nnU-Net deep learning framework. Trained with a total of 1192 [18F]FDG-PET/CT scans from 716 patients with Non-Hodgkin lymphoma participating in the PETAL trial.
For installation and usage instructions, please visit https://github.com/hzdr-MedImaging/LyROI
Please cite nnU-Net and the respective paper when using LyROI.
List of available models:
LyROI_Orig.zip: regular U-Net
LyROI_ResM.zip: residual encoder U-Net (medium)
LyROI_ResL.zip: residual encoder U-Net (large)