How does AI assist with false alarm reduction in video and alarm monitoring?

How can false alarms be minimised in a video/alarm monitoring system? It all depends on the right nuisance alarm reduction software. DeepAlert enhances your operations with a real-time alerting solution that makes use of deep learning AI technology to perform video analytics on a continuous basis – ideal for off-site monitoring.

• By analysing what objects are causing motion in a scene, the system learns to deliver relevant alert notifications in line with your specific configurations.
• This learning process is constant, increasing in accuracy and building better models based on input received. 
• By adding defined exclusion zones,  nuisance alarms can be further reduced.

How does DeepAlert work?

DeepAlert’s powerful, deep neural technology reduces false alarms by filtering and delivering real-time alerts to your VMS. DeepAlert Hub technology can be used at the edge to manage video streams and perform edge analytics where bandwidth is a concern, or cameras at the edge can be configured for direct communication with DeepAlert.

• Simple drop-in software solution that works with almost any surveillance system or modern camera system, including IR and thermal.

• Cloud-based deep learning analytics engine at its core, with machine learning applied throughout the system. 

DeepAlert reduces false alarms by more than 70% compared to conventional edge analytics.

How does it do it? 

1. Motion triggers are reduced by applying masking to areas within the scene of interest. 

1. Object detection analysis is enhanced by powerful deep neural technology.

1. Customisable alert preferences allow you to set what notifications you want to receive and when.  

Fewer false alarms means your monitoring operations can focus on real threats and rule out nuisance alarms to deliver a more confident, effective armed response. 

What objects can DeepAlert detect?

The DeepAlert system has more than 30 different object detection classes: person, vehicle and animal, with sub-classifications of each which makes the solution exceptionally useful for intrusion detection and perimeter protection.  The PPE detection (hard hats and hi-vis vests) and weapon detection modes are also effective as early warning signals of non-compliance or imminent danger.

What makes DeepAlert effective at reducing false alarms? It’s a system that keeps learning, getting even more accurate at object detection with time and human oversight. 

• While most video analytics software like motion detection and object detection rely solely on rule-based algorithms, sometimes there are simply too many variables in the footage to permit effective noise reduction. 
• That’s where deep neural networks make all the difference in improving the accuracy of false alarm filtration, by classifying images for future reference. 
• DeepAlert’s system already has millions of images that have been classified. It works “out of the box” due to the global model architecture
• Tens of millions of parameters exist in the deep neural network and these are tuned with a model training process ultimately to improve accuracy in classification for the datasets provided. 
• Depending on architecture and the computing that powers the image classification, several million images can be processed in a day – and learnings from this deep analysis on camera feed imagery is applied across the neural network to further increase classification accuracy.  

What are the benefits of DeepAlert for video and alarm monitoring companies?

1.  DeepAlert Premium provides an intelligent, intuitive fully-functional video management system for video monitoring stations.
2. DeepAlert Connect provides a layer of alert filtration and management on an existing video management system of your choice. 
3. No hardware required, depending on deployment selected. Works straight out of the box with no site training required.
4. Scale on demand, at low cost. Extend your ability to handle more sites and cameras, simply pay per camera and add more as necessary. 
5. Live System Health visibility giving you peace of mind that your surveillance system is working optimally and without compromise.
6. More efficient operator management with Clear Standard Operating Procedures per alert for operators to follow, as well as an auditable log of operator actions per alert and individual operator reports per shift.
7. Better Customer Relationship Management with automated Site Activity Reports.

Keen to see how DeepAlert can make a positive impact on your monitoring station’s operational efficiency? We’re ready to show you!