matRadiomics: a Novel Radiomics Freeware!
matRadiomics is a complete radiomics freeware that allows the user to complete the whole radiomics workflow, from image visualization to predictive model implementation. It enables a simplified workflow for medical image processing and analysis.
1) matRadiomics is a freeware that allows the user to carry out the whole radiomics workflow.
— Giovanni Pasini (@research_giovap) November 24, 2022
It was presented for the first time in a research paper published in the Journal of Imaging. @MDPIOpenAccess
Paper available in the comments below.#radiomics https://t.co/ZlMFm7GlN8
Introduction
Currently, the most popular radiomics software, like Slicer3D and LIFEx, stop at the feature extraction level, not allowing to perform feature selection and predictive model bulding within the same software. Therefore, the user either needs to switch to another software or needs to know how to code, in order to continue with the radiomics workflow.
On the other end, matRadiomics, a novel and complete radiomics freeware written in MATLAB and Python, allows the user to perform all the steps of a radiomics workflow:
- Image importing and visualization
- Segmentation
- Feature extraction
- Feature selection
- Predictive model building
Attention
matRadiomics is currently a research product, and it is still under development.
Installation and Functionalities
The installation procedure and the main functionalities of matRadiomics are explained in some video tutorials available on the IBFM-CNR Youtube Channel
- Installation
This video tutorial focuses on the installation procedure of matRadiomics. Currently, it works only on Windows 10/11.
- Import DICOM, Segment and Extract Features
This video tutorial focuses on how to import DICOM files, how to segment images and how to extract features using pyradiomics, for which matRadiomics provides a graphical user interface.
- Access DICOM attributes, Import Segmentations, Feature Selection
This video tutorial focuses on how to access DICOM files attributes, how to import DICOM segmentation and how to perform feature selection.
- Build and Save a Machine Learning Model
This tutorial focuses on how to train a Machine learning model and evaluate its performance (accuracy, AUC, Confusion Matrix)
- Feature harmonization, Manual Segmentation, Switch between segmentations
This tutorial focuses on how to perform feature harmonization (ComBat package), how to perform manual segmentation of images (draw and erase tool) and how to switch between multiple segmentations.