intuVision analytics offer the flexibility to detect, track, and classify objects for a wide range of applications from alerting security for suspicious activity to counting people and vehicles to even counting fish in streams. In this blog we will feature an application where intuVision VA is used to detect suspicious or unusual gestures.
intuVision VA's user trainable object classification, can be used to learn and detect person pose and gestures ranging from aiming a firearm, aggressive stance, fall detection, or as in our sample video, pointing at the sky.
Person pose detection was first implemented as a security application for fall detection in a chemical facility under concern of unsafe conditions for employees such as poisoning from a gas leak. intuVision analytics have proven successful by immediate notification of people falling to the floor on detecting trips and other more minor falls - alerting security officers who can send medical help.
Extending the pose detection to a smart city implementation, intuVision VA can be trained to detect suspicious or unusual stances, pointing or aiming a firearm.
The example below shows two situations where gesture or person behavior is used to detect the situation of interest. In the first part of the video, the gesture detection is configured to alarm on someone pointing at the sky. You can see how the person in the yellow shirt is detected and tracked throughout his time in the scene, whereas other people do not generate an alert. Similarly for the second clip, when a person shows signs of distress and falls, the alarm is generated, whereas the person walking through the scene is ignored.
If you think that intuVision VA can help you or your customers, you can reach us here.