Building AI Models with CERR - cerr/CERR GitHub Wiki

CERR facilitates data pre-processing and export to HDF5 format for training AI-based models, beginning with DICOM or CERR-format files.

Index

• Pipeline
• Usage
• Setting up configuration files
• Deploying AI models

Citation

Iyer, A., LoCastro, E., Apte, A. P., Veeraraghavan, H., & Deasy, J. O. Portable framework to deploy deep learning segmentation models for medical images. bioRxiv. https://www.biorxiv.org/content/10.1101/2021.03.17.435903v1. Published March 19, 2021.

Pipeline

---

Usage

paramFilename = 'path/to/configuration/file';
inputDir = 'path/to/input/files';                    %Directory with DICOM or CERR-format files.
outputDir = 'path/to/output/directory';
prepareSegDataset(paramFilename,inputDir,outputDir);

Reading exported files

% Read exported scans
I = h5read('path/to/scan/file','/scan');
% Read exported masks
M = h5read('path/to/mask/file','/mask');