Large retail stores. Shopping malls. College campuses. Transportation hubs. These are all dynamic environments with hundreds of people walking among them.
Even on a good day, these are complex sites to monitor.
To keep us safe, as well as to protect an organisation’s assets and reputation, security managers are constantly looking for better and more efficient ways to improve a site’s security. Even a trained, skilled eye could miss events in a naturally busy environment like a university campus. With those goals in mind, every security professional should be aware of AI-assisted video monitoring.
What is AI-assisted video monitoring?
AI-assisted video monitoring software uses advanced technologies like computer vision and machine learning.
iCetana is a unique patented product that doesn’t just try to recognise faces or apply canned rules. Once installed, iCetana ‘learns’ and ‘remembers’ what normal conditions are like. Whenever circumstances vary from ‘normal’ and move into ‘abnormal’ territory, the system immediately notifies security operators.
For example, when a person starts displaying suspicious behaviour, the system would recognise that this is not normal. The system will display the footage to an operator. The operator can assess the situation and then decide whether to take action or disregard it.
This is different from traditional rule-based surveillance software. The small set of preprogrammed scenarios limits the effectiveness of rule-based software. In contrast, iCetana compares the current situation to what it has learned to be ‘normal’. It can then identify abnormal activities in real time, and highlight these.
The strengths and limitations of AI-assisted video monitoring
AI-assisted video monitoring has become increasingly popular due to the business value it can generate.
Reduced operator fatigue is one of the biggest benefits of AI-assisted surveillance. AI-assisted monitoring only shows potentially actionable incidents. This focuses attention on events that require human judgment.
None of this is to say that AI-assisted video monitoring is perfect, however. While it has a high detection rate, the technology is still unable to identify some situations. For example, identifying shoplifting still needs a trained human eye and intuition. AI-assisted monitoring software may miss certain shoplifting events as they appear less ‘abnormal’ than others.
AI-assisted monitoring can detect many other types of situations in real-time:
- Suspicious behaviour (loitering, casing a site)
- Unauthorised access (people and vehicles)
- Crowd gathering and dispersal
- Manufacturing process breaches
- Violence and fighting
- Medical emergencies
- Vandalism to fixed assets
- Camera tampering
AI-assisted video monitoring also allows teams to take a more proactive, responsive approach to security. The system notifies operators when the event is still underway. Operators can respond immediately instead of waiting and reacting belatedly. Security personnel can protect assets before damage occurs, saving money.
Finally, another one of the major benefits of AI-assisted surveillance is that it scales up with your needs. You don’t need to install AI-assisted monitoring software for all of your cameras. You can start with a smaller set of cameras first, then ramp up from there. There is no need to make major changes in your security solution. You can easily expand coverage to a large number of cameras (several hundred or more) as needed.
AI-assisted video monitoring complements rather than replaces human judgment. A system may highlight an event that an operator determines to be harmless. AI-assisted video monitoring systems massively reducing the amount of information that operators must deal with. This allows them to make more informed and timely decisions.
These systems generate immediate value for an organisation. Not only will you save money by proactively protecting your assets before any damage occurs, but you will boost productivity and reduce fatigue for your security team.