Repository navigation
dicomqc 0.2.0 introduces the Desktop application for auditing DICOM metadata on Linux, macOS, and Windows. It uses the same Python audit engine as the CLI. dicomqc is read-only: it checks privacy risks and compares source and de-identified datasets, but does not modify DICOM files.
Highlights
- Desktop application with portable
.dicomqcprojects, run history, report previews, and synthetic examples. - Dataset comparison for file completeness and patient pseudonym consistency.
- YAML project policies, UID integrity checks, and scanner/private-tag inventory.
- Interactive offline HTML reports and MultiQC-compatible output.
- On-demand update checks and platform-specific application menus.
Desktop downloads
- Linux x86_64 or ARM64: AppImage.
- macOS Apple Silicon or Intel: DMG.
- Windows x86_64: setup executable.
Each installer includes a SHA-256 checksum file. Python and the audit engine are bundled. The macOS builds are not Apple-notarized; Windows installers are unsigned.
Validation
All five attached installers were built from the annotated v0.2.0 tag at commit 5457c4c56cebbca3e5dbaa57ec52fcd19781e245. The five-platform installer workflow passed, including engine, frontend, native bridge, packaging, startup, and checksum checks.
Installer candidates were manually validated on Linux, macOS, and Windows before the final tagged rebuild. The Python package is also available from PyPI.
Audits do not inspect pixels, burned-in text, or facial features and do not certify de-identification or regulatory compliance.
See the installation guide, documentation, and changelog.