PredCRD

Predict Continuous Representation of Disease

Introduction

The pipeline aims to combine knowledge distillation (KD) methods for disease subtyping and representation methods.

Supported Pipelines

Installation

git clone https://github.com/CBICA/PredCRD.git
cd PredCRD/
pip install -e .

DLMUSE ROI volumes to SurrealGAN R-indices pipeline:

EXAMPLE USAGE:
PredCRD -i /path/to/input.csv
-o /path/to/output.csv
-d *Optional cuda/cpu
-m *Optional /path/to/model.pth
-s *Optional /path/to/scalar.bin
-mt *Optional /path/to/mean_icv.npy

*note: default model.pth, scalar.bin, and mean_icv.npy are included in model/ folder