Platform

MaskAnyone Toolkit

Privacy-aware audiovisual data management infrastructure. De-identify video and audio while preserving research-relevant features.

Get Started View on GitHub
Capabilities

Platform Features

Face De-identification

Multiple masking methods including face swapping, blurring, and pixelation while preserving facial expressions.

Face Swap Blur Pixelate

Voice Anonymization

Protect speaker identity through pitch shifting and formant modification while maintaining prosodic patterns.

Pitch Shift Formant

Body Movement Tracking

Extract skeletal pose data and gesture information for behavioral analysis research.

Pose Gesture Kinematics

Cloud Infrastructure

Secure cloud-based processing via SURF infrastructure with encrypted storage and GDPR compliance.

SURF Encrypted

FAIR Data Archiving

Integration with DANS for long-term archiving following FAIR principles.

DANS DOI

User-Friendly Interface

Point-and-click web interface requiring no programming knowledge.

Web UI Presets
Process

Research Workflow

Upload Data

Securely upload audiovisual research data via encrypted connection. Supports multiple video formats and batch processing.

Configure Masking

Choose from predefined presets or create custom configurations. Preview masking effects before processing.

Process & Extract

Automated detection and masking. Simultaneous extraction of pose data, gestures, and behavioral features.

Quality Check

Review processed videos and extracted data. Verify privacy protection while ensuring research utility.

Archive & Share

Export to DANS for long-term archiving or share via secure links. Generate metadata and DOIs.

Key Benefits

Privacy Protection

GDPR-compliant de-identification enabling open science practices.

Data Reusability

Enriched data with behavioral annotations increases value for secondary analysis.

Accessible to All

No programming skills required. Point-and-click interface.

Reproducible Research

Documented workflows and version control ensure transparency.

Architecture

Technical Stack

System Architecture

SYNAPSIS is built on a modular, cloud-native architecture separating the user interface, processing engine, and storage layers.

Frontend

React-based web interface with real-time preview

Processing

Python/PyTorch backend with GPU acceleration

Storage

SURF SANE secure storage with DANS archiving

Masking Techniques

Face Masking Methods

Face Swap (SimSwap)

Replace face with synthetic identity while preserving expressions

High Privacy Preserves Expression
Gaussian Blur

Traditional blurring with adjustable intensity

Fast Standard
Pixelation

Reduce resolution in facial region

Simple Effective
DeepPrivacy2

GAN-based realistic face synthesis

Realistic State-of-art

Audio Masking Methods

Pitch Shifting

Modify voice pitch while maintaining prosody

Formant Modification

Change vocal tract characteristics

Infrastructure & Security

Security & Compliance

GDPR Compliant
End-to-End Encryption
Role-based Access Control
SURF Infrastructure
University HPC Clusters (Ponyland)

Performance

~5min
per minute of video
GPU
accelerated
Batch
processing
4K
max resolution

Infrastructure & Partners

Ponyland HPC

Radboud GPU cluster

SURF SANE

Cloud computing

DANS

Data archiving

CLARIAH

Integration

API & Integration

For advanced users, SYNAPSIS provides a REST API for programmatic access to masking capabilities.

# Example Python API usage
import synapsis

# Initialize client
client = synapsis.Client(api_key="your_api_key")

# Upload and process video
video = client.upload("path/to/video.mp4")
result = video.mask(
    face_method="face_swap",
    audio_method="pitch_shift",
    extract_pose=True
)

# Download results
result.download_video("masked_output.mp4")
result.download_pose_data("pose_data.json")

API Access Requirements

API access requires institutional affiliation and completed training. Apply through your data steward.

Live Demo

MaskAnyone Platform

MaskAnyone is deployed on Radboud University's Ponyland HPC cluster with GPU acceleration. Access is currently restricted to project partners and pilot participants.

Request Access View on Codeberg

Hosted on Ponyland (Radboud University HPC). Contact us to arrange a demo session.

Start

Get Started with SYNAPSIS

1

Register

Create an account using your institutional email address.

Coming Soon
2

Get Trained

Complete the Masking School introduction course.

View Courses
3

Start Masking

Access the platform and begin processing your data.

Coming Soon

System Requirements

Web Interface

  • Modern web browser
  • Stable internet connection
  • Institutional network access

Local Installation (Optional)

  • Python 3.8+
  • CUDA-capable GPU (recommended)
  • 16GB RAM minimum