When most people think about cameras in aged care, they think about surveillance. But a new generation of computer vision technology is doing something fundamentally different: it's watching for safety, not watching people. And the distinction matters enormously.
Computer vision - the branch of AI that enables machines to interpret visual information - is quietly becoming one of the most impactful technologies in aged care. Here's how it's being used, and why providers should be paying attention.
Falls remain the leading cause of injury-related hospitalisation for older Australians. Traditional fall detection relied on wearable pendants that residents had to press - a system that fails precisely when it's needed most, since many falls result in confusion, unconsciousness, or an inability to reach the button.
Computer vision-based fall detection uses cameras or depth sensors to identify fall events automatically. Modern systems can distinguish a genuine fall from someone bending down to pick something up, sitting down quickly, or even a blanket sliding off a bed. The AI has been trained on thousands of real scenarios to reduce false alarms to near zero while catching genuine incidents within seconds.
For residential care providers, this means faster response times, fewer unnecessary callouts, and - critically - detection of falls that would otherwise go unnoticed until the next check. For home care, camera-free alternatives using radar and depth sensors offer the same protection without the privacy concerns of traditional video.
One of the most fascinating applications comes from Australian startup Lookinglass, which is developing a smart mirror that uses computer vision and AI to detect early signs of conditions like Parkinson's disease and dementia.
The concept is elegantly simple: a resident looks in the mirror as part of their normal routine. The system analyses subtle facial movements, micro-expressions, and physical indicators that the human eye would miss. Changes in facial muscle control, eye movement patterns, and postural shifts can signal neurological changes months or even years before traditional diagnosis.
This isn't about replacing clinical assessment. It's about flagging concerns earlier, when interventions are most effective and care plans can be adjusted proactively rather than reactively.
Beyond individual events like falls, computer vision is being used to understand patterns of behaviour over time. Systems can learn a resident's typical movement patterns - when they usually get up, how they navigate common areas, their gait speed and steadiness - and alert staff when something changes.
A resident who normally walks steadily but has started shuffling. Someone who usually visits the dining room independently but has been hesitating at doorways. A person whose nighttime movement patterns have changed significantly. These subtle shifts often precede significant health events, and computer vision can detect them continuously in a way that periodic human observation cannot.
The elephant in the room with any visual monitoring technology is privacy. Aged care residents deserve dignity and autonomy, and the sector rightly takes a cautious approach to surveillance.
The good news is that modern computer vision systems are designed with privacy at their core. Many systems process video locally on the device and only transmit anonymised data - skeleton outlines, movement patterns, alert events - rather than actual footage. Some use thermal or depth sensors instead of cameras entirely, capturing spatial data without ever recording a recognisable image.
The key for providers is choosing systems that are transparent about how data is captured, processed, and stored, and that give residents and families genuine control over their participation.
Computer vision in aged care is mature enough to deploy but still early enough that adopting it now puts you ahead. If you're considering it for your organisation, start with the use case that addresses your most pressing safety concern - whether that's falls, wandering, or early deterioration detection - and look for providers who have real Australian deployment experience, not just international case studies.
The technology is here to make care safer, not more intrusive. And for providers who get the implementation right, it's a powerful differentiator in an increasingly competitive market.
Generation Tech. Generation Care.