search

biomedia-mira / stochastic_segmentation_networks

Stochastic Segmentation Networks
MIT License
65 stars 10 forks source link

pytorch code for LIDC dataset #6

Closed ZhouCX117 closed 2 years ago

ZhouCX117 commented 2 years ago

Hi, thanks for your excellent work! Is there a PyTorch code for the LIDC dataset? I could not fully understand the paper, so I want to run the code to help to understand. The PyTorch code for the 3D dataset is a little hard for me and I couldn't download the dataset. Looking award to your answer!

MiguelMonteiro commented 2 years ago

Hi, the code for the LIDC dataset is in tensorflow as it is a fork form the PHISEG code: LIDC 2D. See the bottom of the readme:

The LIDC 2D results presented in the paper were generated using a fork of the PHi-Seg code. Evaluation script is phiseg_full_evaluation.py. Loss function in phiseg/phiseg_model.py. Model in phiseg/model_zoo/likelihoods.py. Additional configs in phiseg/experiments. You need to modify config/system.py, phiseg_full_evaluation.py and phiseg/experiments/* with your own local paths.