i-PRO Co. Ltd., a global pioneer in professional security solutions for surveillance and public safety, will add AI Scene Change Detection to its collection of powerful edge-based AI analytics applications.
AI-based analytic application
Embodying the latest in AI analytics technology, Scene Change Detection is the first AI-based analytic application that provides easy on-site learning of a scene. When anomalies occur within the camera’s field of view, an alert can be sent to security and operations teams. The new application will be deployable to more than 100 i-PRO AI camera models upon release.
Unlike traditional AI-based object detection analytics which alert or flag when objects cross lines or enter predefined zones, scene change detection allows users to teach a surveillance camera what an entire scene looks like in a normal state.
Customised threshold settings
Using AI-based object detection ensures that people or vehicles that pass through the scene can be automatically excluded
The analytic then flag an anomaly whenever any preselected part of the scene varies from normal. This might include a door propped open longer than normal or stock on a store shelf that sells beyond a preset limit.
Because of the customised threshold settings, brief changes in the scene are ignored since the scene returns to normal within an acceptable timeframe. Utilising AI-based object detection ensures that people or vehicles that pass through the scene can be automatically excluded.
AI-based object detection app
“We believe that i-PRO’s Scene Change Detection represents the industry’s first edge AI-based scene change detection app to optimise the scene detection algorithm by learning from captured images,” says Norio Hitsuishi, Head of Product Management at i-PRO Co., Ltd.
“Our unique implementation is very powerful and allows users to specify how much change triggers an alert after a specified time."
Scene change detection
Norio Hitsuishi adds, "For example, an outside door should be able to open and close routinely without triggering an alarm. Using Scene Change Detection, users can set it to only alarm if the door is left open for more than 5 minutes. This is something that traditional motion detection analytics can’t do.”