Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Meta-Guardian: An Early Evaluation of an On-device Application to Mitigate Psychography Data Leakage in Immersive Technologies

Keshav Sood, Sanjay Selvaraj, Youyang Qu

TL;DR

This work introduces Meta-Guardian, an on-device privacy-preserving framework and Unity SDK that detects and filters biometric signals within VR headsets in real time to prevent leakage of sensitive data. It deploys a lightweight $14$-dimensional feature classifier on the Meta Quest Pro using ONNX-exported models run via Unity Sentis, enforcing privacy by design before any data leaves the device. Experimental evaluation across three XR environments demonstrates that privacy filtering preserves immersive feedback and maintains classifier performance, with minimal impact on user interaction. The approach offers a practical, production-ready solution for developers to mitigate biometric privacy risks in immersive technologies and supports regulatory compliance through on-device data minimization.

Abstract

The use of Immersive Technologies has shown its potential to revolutionize many sectors such as health, entertainment, education, and industrial sectors. Immersive technologies such as Virtual Reality (VR), Augmented reality (AR), and Mixed Reality (MR) have redefined user interaction through real-time biometric and behavioral tracking. Although Immersive Technologies (XR) essentially need the collection of the biometric data which acts as a baseline to create immersive experience, however, this ongoing feedback information (includes biometrics) poses critical privacy concerns due to the sensitive nature of the data collected. A comprehensive review of recent literature explored the technical dimensions of related problem; however, they largely overlook the challenge particularly the intricacies of real-time biometric data filtering within head-mounted display system. Motivated from this, in this work, we propose a novel privacy-preserving system architecture that identifies and filters biometric signals (within the VR headset) in real-time before transmission or storage. Implemented as a modular Unity Software-development Kit (SDK) compatible with major immersive platforms, our solution (named Meta-Guardian) employs machine learning models for signal classification and a filtering mechanism to block sensitive data. This framework aims to enable developers to embed privacy-by-design principles into immersive experiences on various headsets and applications.

Meta-Guardian: An Early Evaluation of an On-device Application to Mitigate Psychography Data Leakage in Immersive Technologies

TL;DR

This work introduces Meta-Guardian, an on-device privacy-preserving framework and Unity SDK that detects and filters biometric signals within VR headsets in real time to prevent leakage of sensitive data. It deploys a lightweight -dimensional feature classifier on the Meta Quest Pro using ONNX-exported models run via Unity Sentis, enforcing privacy by design before any data leaves the device. Experimental evaluation across three XR environments demonstrates that privacy filtering preserves immersive feedback and maintains classifier performance, with minimal impact on user interaction. The approach offers a practical, production-ready solution for developers to mitigate biometric privacy risks in immersive technologies and supports regulatory compliance through on-device data minimization.

Abstract

The use of Immersive Technologies has shown its potential to revolutionize many sectors such as health, entertainment, education, and industrial sectors. Immersive technologies such as Virtual Reality (VR), Augmented reality (AR), and Mixed Reality (MR) have redefined user interaction through real-time biometric and behavioral tracking. Although Immersive Technologies (XR) essentially need the collection of the biometric data which acts as a baseline to create immersive experience, however, this ongoing feedback information (includes biometrics) poses critical privacy concerns due to the sensitive nature of the data collected. A comprehensive review of recent literature explored the technical dimensions of related problem; however, they largely overlook the challenge particularly the intricacies of real-time biometric data filtering within head-mounted display system. Motivated from this, in this work, we propose a novel privacy-preserving system architecture that identifies and filters biometric signals (within the VR headset) in real-time before transmission or storage. Implemented as a modular Unity Software-development Kit (SDK) compatible with major immersive platforms, our solution (named Meta-Guardian) employs machine learning models for signal classification and a filtering mechanism to block sensitive data. This framework aims to enable developers to embed privacy-by-design principles into immersive experiences on various headsets and applications.
Paper Structure (24 sections, 8 figures, 5 tables)

This paper contains 24 sections, 8 figures, 5 tables.

Table of Contents

  1. Introduction and Background
  2. Introduction
  3. Related Work
  4. Meta-Guardian: Methodology and Design
  5. Biometric Filtering and Privacy Enforcement
  6. Development of the proposed SDK: Implementation Guide:
  7. On VR Device and Platform Rationale
  8. Threat Model
  9. Experimental Design and Data Collection
  10. Classifier Architecture and Training
  11. Architecture and Model Design
  12. Data Collection and Labeling Strategy
  13. Deployment Pipeline: From PyTorch to Unity
  14. Unity Sentis Inference Code
  15. Feature Summary and Interpretability
  16. ...and 9 more sections

Figures (8)

  • Figure 1: Framework Architecture
  • Figure 2: The Meta Quest Pro headset and Touch Pro controllers, showing the inward-facing IR cameras and sleek mixed-reality form factor.
  • Figure 3: Biometric Classification Flow.
  • Figure 4: Annotated 3D visualization of facial expression weights captured from a representative XR session. Highlighted muscle regions correspond to actively engaged facial features, with associated intensity values (range: 0–1) indicating the degree of activation. This snapshot reveals moderate brow tension, subtle cheek and lip movement, and distinct gaze orientation demonstrating how facial data can encode cognitive and emotional cues relevant to biometric inference.
  • Figure 5: a) Environment 1 – Expression Graph (Pre SDK), b) Environment 1 – Expression Graph (Post SDK)
  • ...and 3 more figures