[4aad23]: / installation / README.md

Download this file

80 lines (62 with data), 2.5 kB

Installation

Table of Contents

Getting Started

Install the requirements:

  • Linux or macOS
  • Python 3
  • CPU or NVIDIA GPU + CUDA CuDNN

  • Clone this repo:

git clone -b master --single-branch https://github.com/moralesq/DeepStrain.git
cd DeepStrain

For pip users:

bash ./scripts/install_pip.sh

Pre-trained Models

  • Download some test data in nifti format:
bash ./datasets/download_sample_dataset.sh
  • Download the pre-trained models (i.e., carson and carmen):
bash ./pretrained_models/download_model.sh

Automated Scripts

Segmentation

  • Generate segmentations with the model for 3D niftis:
bash ./scripts/test_segmentation.sh ./datasets/sample_nifti_3D NIFTI ./results/sample_nifti_3D

The test results will be saved to a nifti file here: ./results/sample_nifti_3D/.

  • Generate segmentations with the model for 4D (3D+time) niftis:
bash ./scripts/test_segmentation.sh ./datasets/sample_nifti_4D NIFTI ./results/sample_nifti_3D

The test results will be saved to a nifti file here: ./results/sample_nifti_4D/.

Motion Estimation

  • Generate motion estimates with the model for 4D (3D+time) niftis. Note that for motion estimation only 4D is supported:
bash ./scripts/test_motion.sh ./datasets/sample_nifti_4D NIFTI ./results/sample_nifti_4D

The test results will be saved to a h5 file here: ./results/sample_nifti_4D/.

  • Generate both segmentation and motion estimates with the model for 4D (3D+time) niftis:
bash ./scripts/test_segmentation_motion.sh ./datasets/sample_nifti_4D NIFTI ./results/sample_nifti_4D

The test results will be saved to nifti and h5 files here: ./results/sample_nifti_4D/.

Strain Analysis

  • After the segmentations and motion estimates have been generated, we can use both to calculate myocardial strain. Note that we're passing the output folder from the previous runs:
bash ./scripts/test_strain.sh ./results/sample_nifti_4D

The test results (4D radial and circumferential strain) will be saved to nifti files here: ./results/sample_nifti_4D/.