clinicadl Documentation

Installation

clinicadl can be installed on Mac OS X and Linux machines, and possibly on Windows computers with a Linux Virtual Machine.

We assume that users installing and using clinicadl are comfortable with using the command line.

• Installation

User documentation (clinicadl)

Prepare your imaging data

• clinicadl preprocessing run - Preprocessing pipelines
  • t1-linear - Linear processing of T1w MR images: affine registration to the MNI standard space
  • t1-extensive - 'Extensive' processing of T1w MR images: non linear registration to the MNI standard space
• clinicadl preprocessing quality-check - Quality control of preprocessed data: use a pretrained network [Fonov et al., 2018] to classify adequately registered images.
• clinicadl preprocessing extract-tensor - Prepare input data for deep learning with PyTorch

Train & test your classifier

• clinicadl random-search - Explore hyperparameters space by training random models
• clinicadl train - Train with your data and create a model
• clinicadl classify - Classify one image or a list of images with your previously trained CNN
• clinicadl interpret- Interpret trained CNNs on individual or group of images

Utilitaries

• clinicadl generate - Generate synthetic data for functional tests
• clinicadl tsvtool - Handle TSV files for metadata processing and data splits

Pretrained models

Pretrained models for the CNN networks implemented in ClinicaDL can be obtained here: https://zenodo.org/record/3491003

These models were obtained during the experiments for publication. They correspond to a previous version of ClinicaDL, hence their file system is not compatible with the current version. Updated versions of the models will be published soon.

Support

• Report an issue on GitHub
• Use the clinicadl Google Group to ask for help!

License

clinicadl is distributed under the terms of the MIT license given here.

Citing clinicadl

For publications or communications using clinicadl, please cite [Wen et al., 2020] as well as the references mentioned on the wiki page of the pipelines you used (for example, citing PyTorch when using the extract pipeline).

Disclaimer

clinicadl is a software for research studies. It is not intended for use in medical routine.

---