VLM output
A thermal (infrared) camera monitors a single person sitting on the edge of a bed in a dark room, who leans forward and reaches toward a bedside table or rail. Over the frames the person moves progressively lower and further to the right, ending face-down and slumped over the edge of the bed with their body partly out of the bed area, with no other person present.
- Fall Risk DetectedYes
- Patient Out Of BedPartial
- Bed Exit AttemptYes
- Person Count1
- Caregiver PresentNo
- PostureSlumped Forward
- Unattended PatientYes
- Alert LevelHigh
- Lighting ConditionThermal/Dark
Customer challenges
Patient falls are a leading cause of injury and liability and often occur unwitnessed, yet constant staff observation isn't feasible; patients may be unable to reach a call button, and staff can't watch every room continuously; and fire in a hospital is high-risk with vulnerable, hard-to-evacuate patients, while point detectors can be slow in large or compartmentalized spaces.
Customer benefits
Faster response with reduced injury severity, fewer fall-related complications and claims, and coverage of unwitnessed areas; a safety net beyond call buttons with wider coverage and no extra staff; and earlier fire warning for safer evacuation, protecting patients and assets — all using existing cameras.
Key features
Incident Detection — Person Fall
Detects a person falling and raises an incident alert.
Distress Monitoring (waving)
Detects people waving for help or showing signs of distress.
Fire & Smoke Detection
Detects fire and smoke early in monitored areas.
Before & after
| Before | After — with Linker AI Nexa |
|---|---|
| Before Falls discovered only on rounds or when a patient calls out, with delayed response and unclear cause; distress noticed only on rounds or if a patient reaches the call button; and fire detected only after smoke reaches ceiling sensors or is seen by staff. | After — with Linker AI Nexa A fall triggers an automatic alert to the nursing station the moment it happens, with a timestamped clip; distress gestures are flagged live to staff for immediate attention; and smoke or flame is flagged at the source within seconds, with location pinpointed. |