FAQ

Radiuma is a user-friendly, open-source software platform designed to help medical professionals and researchers visualize, process, and analyze medical images for radiomics studies.

Radiuma is compatible with both Windows and Unix/macOS environments. It requires Python 3.11.4 and includes automated installation scripts to ensure a smooth setup on all these systems.

Radiuma offers broad compatibility with a wide range of standard medical imaging and data formats, ensuring seamless integration into your research workflow:

Medical Images: Full support for NIFTI (.nii, .nii.gz), DICOM (individual files and complete studies), DICOM-RT (for radiotherapy structures), and NRRD formats.
Imaging Modalities: Compatible with all major modalities, including CT, MRI, PET, SPECT, and Ultrasound.
Tabular Data: Easy import/export for .csv and .xlsx files for statistical analysis.

No. Radiuma is designed for various expertise levels. Whether you are a radiation oncologist, radiologist, medical physicist, or data scientist, the visual node-based system makes it easy to build and run workflows without advanced programming skills.

PySERA (Python-based Standardized Extraction for Radiomics Analysis) is a high-performance, open-source library that serves as a bridge between traditional radiomics and modern deep learning. It provides a robust, Python-native environment for extracting quantitative biomarkers from medical images with a rigorous focus on scientific reproducibility.

It automates the entire radiomics pipeline, from standardized preprocessing (resampling, discretization, and normalization) to feature extraction. It computes over 550 features—including IBSI-compliant handcrafted features and advanced deep learning radiomics embeddings—using a parallelized engine optimized for large-scale, multi-format medical imaging datasets.

AllMetrics is a unified, domain-agnostic Python library engineered to standardize performance evaluation in machine learning. It serves as a central hub for consistent metric computation, eliminating the discrepancies often found when comparing results across different programming environments (Python, MATLAB, R) or disparate software frameworks.

It provides a modular API that handles everything from basic regression and classification metrics to complex segmentation and image-to-image translation evaluation. Beyond simple computation, it incorporates automated data validation protocols to detect and resolve reporting discrepancies before they impact your model’s final assessment.