AI-Powered Patient Safety Monitoring That Protects Privacy
The system uses DeCloak's AipA (AI-aided Privacy Agent) technology to provide real-time patient movement tracking and critical event detection across healthcare facility environments. Camera feeds are processed by on-device AI to detect falls, incidents, and unusual movement patterns — while simultaneously de-identifying patient imagery before any data is stored or transmitted.
Unlike conventional CCTV surveillance, the system does not retain identifiable footage of patients. Privacy-preserving processing occurs at the point of capture, enabling healthcare facilities to monitor patient safety continuously without creating a database of patient imagery.
This product is not a medical device and is not TGA-registered. It is for patient safety purposes only.
Real-time de-identification at point of capture — no identifiable patient imagery stored or transmitted.
Automatic detection of patient falls with immediate staff alert notification.
AI recognition of incidents, agitation, and medical emergencies requiring rapid response.
For patient safety purposes only — no TGA registration, no clinical diagnostic claims.
How Patient Movement Monitoring Works
Privacy-preserving AI processes camera feeds in real time — detecting events and alerting staff without retaining patient imagery.
Camera Capture
Facility cameras capture movement in patient areas, corridors, and wards. No specialised hardware required — integrates with existing camera infrastructure.
Real-Time De-identification
DeCloak's AI processes video at the edge or local server before storage — removing personally identifiable features from all imagery.
AI Event Detection
Multimodal deep neural networks analyse de-identified movement data to detect falls, unusual inactivity, agitation, and critical safety incidents.
Instant Staff Alert
Detected events trigger immediate alerts to nursing stations, mobile devices, or nurse call systems — enabling faster response.
Safety Reporting
De-identified event logs available for facility safety review, incident reporting, and quality improvement — without retaining patient-identifiable records.
Built for Healthcare Safety Environments
Designed to improve patient safety outcomes while meeting the privacy expectations of healthcare facility governance and patients.
Privacy-Preserving De-identification
Patented AI de-identification removes identifiable patient features from imagery in real time at point of capture. No identifiable patient footage is retained in storage or transmitted across networks.
Fall Detection & Alert
Automatic detection of patient falls using AI movement analysis. Detected falls trigger immediate alerts to nursing staff, reducing the time between incident and response.
Critical Event Detection
AI recognition of aggressive behaviour, medical emergencies, and critical safety events — including detection of extended inactivity and unusual movement patterns.
24/7 Continuous Monitoring
Automated monitoring operates continuously without requiring dedicated observation staff. Reduces nursing workload associated with manual patient observation.
Existing Infrastructure
Designed to work with existing facility camera systems, nurse call platforms, and local network environments — reducing deployment cost and complexity.
De-identified Safety Reporting
Event data logged in de-identified form, supporting facility incident reporting, safety committee review, and quality improvement without privacy obligations.
Where It Can Be Deployed
Suitable for any healthcare environment where patient safety monitoring, fall prevention, and rapid staff response are operational priorities.
Acute Care Hospital Wards
Continuous monitoring of general medical and surgical wards where patient mobility, fall risk, and rapid deterioration are ongoing safety concerns.
Aged Care & Residential Facilities
Monitoring of residents where fall risk is elevated and staffing ratios may limit one-to-one observation. Particularly relevant for dementia care units.
Mental Health Inpatient Units
Detection of critical safety incidents in psychiatric inpatient settings — including self-harm risk monitoring and agitation detection — while preserving patient dignity.
Emergency Departments
Monitoring of waiting areas and treatment bays where patient deterioration, agitation, and fall risk require rapid response across high-volume environments.
Intensive Care Units
Supplementary monitoring of ICU patients where extended monitoring periods, sedation, and mobility restrictions create specific fall and incident detection requirements.
Rehabilitation Wards
Monitoring of patients undergoing physical rehabilitation where supervised mobility, fall risk during ambulation, and activity tracking are patient safety priorities.
Clinical & Industry Validation
DeCloak's patient safety AI has been deployed in real-world healthcare environments and presented at leading AI healthcare conferences.
National Taiwan University Hospital — Hsin-Chu Branch
DeCloak's AipA Healthcare System is operational at National Taiwan University Hospital (NTUH) Hsin-Chu Branch Zhubei Campus — one of Taiwan's leading tertiary referral hospitals. The system operates across both clinical wards and public areas, providing real-time patient safety monitoring with privacy-preserving de-identification.
AI Healthcare Presentation: Multimodal DNN and VLM for Event Detection
Dr. Yao-Tung Tsou, President of DeCloak Intelligences, presented at NVIDIA GTC 2025 on "Enhance Patient Safety With Privacy-Preserving AI: Multimodal DNN and VLM for Event Detection in Healthcare." The presentation demonstrated the AipA system's capabilities in real-world hospital environments in partnership with NTUH Hsin-Chu Branch.
Differential Privacy & Homomorphic Encryption
The privacy-preserving architecture is built on differential privacy techniques and homomorphic encryption — enabling AI analysis on de-identified data without exposing underlying patient information. Designed to support compliance with GDPR and Australia's Privacy Act 1988.
Not a Medical Device
The system is not a medical device under the Therapeutic Goods Act 1989 (Cth) and is not listed on the ARTG. It does not diagnose, prevent, or treat any medical condition. The system is a patient safety monitoring tool for healthcare facility operations and safety staff only.
Common Questions
Contact us for facility-specific deployment enquiries.
Product Documents
Technical and product documentation for facility evaluation and procurement. Contact us for additional materials.
For patient safety purposes only. The Patient Movement Monitoring system is not a medical device under the Therapeutic Goods Act 1989 (Cth) and is not listed on the Australian Register of Therapeutic Goods (ARTG). It is not a TGA-registered in vitro diagnostic device. The system does not diagnose, prevent, monitor, or treat any medical condition and makes no clinical diagnostic claims. It is intended solely for use by healthcare facility operations and safety staff as a patient safety monitoring tool. System outputs must not be used as the basis for clinical diagnostic or treatment decisions. Refer to the product documentation for full intended use and limitations. Supplied in Australia by Advanced Biotech Pty Ltd.
Enquire About Patient Movement Monitoring
Contact Advanced Biotech Pty Ltd for product information, facility demonstrations, pricing, and deployment enquiries across Australia.