IsloPol
Circular Economy2025Turning EPS waste into local circular value.
Back to projects
Protecting identities in XR and video conferencing through multimodal anonymization.
XR-Care
XR-Care is a privacy-preserving solution that automatically anonymizes recorded XR and video-conferencing sessions. Designed for professional settings such as healthcare, education, and corporate communication, XR-Care removes identifiable audio, text, and visual information while preserving the usefulness of the content. As remote communication increases, so does the risk of sensitive information appearing in recorded sessions—faces, voices, documents, screen text, ID badges, and location clues. XR-Care ensures that this information is securely anonymized, enabling organizations to safely store, review, and analyze recordings without exposing personal data or violating GDPR.
Remote consultations, virtual meetings, and XR-based interactions increasingly contain highly sensitive personal information. Faces, voices, screen text, and background details can inadvertently reveal the identity, location, or private context of participants. Yet most conferencing platforms simply record everything as-is, offering no native anonymization tools and leaving organizations exposed to privacy risks and regulatory pressure.
For institutions in healthcare, education, and corporate environments, this creates a major obstacle. Valuable knowledge is lost because recorded sessions cannot be safely archived, reviewed, or repurposed. Without a secure approach to anonymization, organizations face strict GDPR limitations, making it difficult to use these recordings for training, assessment, or long-term research.
Comprehensive Multimodal Protection
XR-Care anonymizes visuals, audio, and text simultaneously. Faces are blurred or masked, sensitive text is detected and obscured, and voices are transformed into non-identifiable speech while retaining meaning.
GDPR-Compliant Recording Workflows
All identifying information is removed before storage, ensuring full compliance with GDPR and sector-specific regulations. Organizations can confidently archive and reuse sessions without risking exposure of personal data.
XR-Care makes recorded sessions safe, usable, and compliant — unlocking new value from data that was previously too sensitive to store or share.
XR-Care integrates seamlessly with existing XR and video conferencing platforms. After a session is recorded, the raw audio, video, and screen content are automatically processed through XR-Care’s anonymization pipeline. The system identifies faces, documents, badges, and contextual clues on-screen, applying blurring or masking techniques to remove any identifiable visual information. At the same time, displayed text is detected and anonymized, ensuring that names, diagnoses, tags, or internal notes do not appear in the final recording.
Audio undergoes its own anonymization stage. Voices are transformed using pitch shifting and spectral modifications that remove identity but preserve conversational clarity. Once all sensitive data has been removed, the anonymized recording is securely stored in compliance with GDPR, ready for training, documentation, research, or quality review—without exposing personal information.
Stronger Security for Sensitive Digital Communication
XR-Care streamlines privacy protection by automating anonymization processes that would otherwise be slow, manual, and inconsistent. Organizations gain secure, reusable recordings that maintain their value for training, review, and long-term analysis.
99%
Reduction in identifiable visual data
90%
Accuracy in detecting on-screen text
80%
Voice identity removal
100%
GDPR-compliant storage
Multimodal anonymization allows organizations to use real recordings without putting personal information at risk.
XR-Care dashboard overview
Track projects, original videos, anonymized videos and frames in real time.
XR-Care sets a new standard for privacy-preserving communication. It enables real-world training and analysis using actual sessions—without exposing any personal information.
CORTEX² Project Partner
