About

About MaskAnyone

Privacy-preserving video anonymization for research. Part of the SYNAPSIS project, funded by NWO (Dutch Research Council).

What is MaskAnyone?

MaskAnyone is an open-source tool for anonymizing people in video recordings while preserving research-relevant features like gestures, body language, and spatial relationships. It's designed specifically for Social Sciences and Humanities (SSH) researchers who need to share audiovisual data while protecting participant privacy.

Multi-Person Tracking

Automatically detect and track multiple people throughout your video with consistent IDs.

Consent-Aware

Set different consent levels per person. Full consent, partial consent, or full anonymization.

Flexible Methods

Choose from blur, pixelation, solid masks, or skeleton overlays based on your research needs.

Batch Processing

Process entire research corpora with consistent settings across all recordings.

Quality Evaluation with MaskBench

How do you know if your anonymization preserves what matters for your research? MaskBench is a companion tool that helps you evaluate anonymization quality across multiple dimensions.

  • Re-identification Risk

    Measure how well faces are protected against re-identification using state-of-the-art face recognition models.

  • Gesture Preservation

    Verify that hand movements and body gestures remain analyzable after anonymization.

  • Emotion Detectability

    Test whether emotional expressions can still be detected or are appropriately obscured.

View MaskBench on GitHub

Archiving Your Research Data

After anonymizing your videos, consider archiving them for long-term preservation, reproducibility, and potential reuse by other researchers. Proper archiving also helps you meet funder requirements and institutional data management policies.

Best Practices for Archiving Anonymized Video

Before depositing data, check with your institution's research data management team about approved repositories and workflows. Ensure your data management plan (DMP) and ethics approval cover data sharing. MaskAnyone can generate processing reports to document your anonymization methods for transparency.

DANS (Netherlands) Zenodo (CERN) Dataverse UK Data Service

The SYNAPSIS Project

MaskAnyone is part of SYNAPSIS (SYstem for privacy-aware aNAlysis of audioviSual research data Including multi-modal tranScriptions), an NWO-funded project to develop privacy-preserving infrastructure for audiovisual research data in the Social Sciences and Humanities.

The project brings together expertise from:

  • Privacy-preserving machine learning
  • Human-centered AI design
  • Research data management
  • SSH research methodology
Visit SYNAPSIS Project Website

Open Source

MaskAnyone is open source software. You can inspect the code, contribute improvements, or deploy your own instance. We welcome contributions from the research community.

View on GitHub Report an Issue

Help & Support