AI Video Analysis on the Edge
- The advanced artificial intelligence human recognition engine can greatly reduce the adverse effects caused by small animal distraction, light rain, light changes, etc.
- Higher accuracy and fewer false positives.
- Various event rules for different application scenarios.
- Video Content Analysis (VCA) detectors for intrusion, loitering, missing object, unattended object, line crossing, and face detection.
Jarvis Smart VCA is a next-generation video analytics suite running on Jarvis cameras for security applications.
The integrated Smart VCA suite includes intrusion detection, loitering detection, missing object detection, unattended object detection, line crossing detection, and face detection.
Deep learning technology
At the core of Jarvis Smart VCA is a deep learning artificial intelligence engine, which can provide higher accuracy of people detection than traditional rule-based design.
Jarvis uses its rich experience in the surveillance field and the resulting huge image database to develop a neural network-assisted human form recognition engine.
Based on the human body contour database and the fast response through the pre-trained deep learning model, the intelligent engine instantly detects the appearance of the human body shape in the monitoring area.
Since most video surveillance is human-centric, the human detection feature allows users to quickly configure their installation. For person detection, only human activities are used as event triggers, while swaying vegetation, passing vehicles, or animal activities in the scene are ignored.
This helps reduce false positives and the time and effort required for camera configuration.
Multiple activity rules
Multiple event rules are deployed on top of the AI core to detect intruders, people loitering or running or crossing lines in an area, lost and abandoned objects, or faces of people in an area. These rules cover most typical security application needs, and multiple rules can be applied to a single camera view.
Intrusion Detection
Jarvis Intrusion Detection can be used to detect people entering or leaving a virtual area within the camera’s field of view.
Area of use:
- Detect when someone enters a bank vault or school after hours.
- Detect when someone leaves the emergency exit and/or fire escape.
Loitering Detection
Jarvis Loitering Detection is used to detect a person or a group of people staying in an area for more than a preset time threshold.
Area of use:
- Detect people standing in walk-through or drive-through ATM lanes.
- Identify high-risk areas of the store where a person is.
Line Crossing Detection
Jarvis Line Crossing Detection detects a person or group of people crossing a virtual trip wire.
Area of use:
- Recognize someone entering a driveway, entrance or exit.
- Identify fence boundaries when:
(1) Someone is near the perimeter
(2) Someone crossed the line Jarvis
Face Detection
Face Detection recognizes and tracks faces in the camera’s field of view.
Area of use:
- Detected and captured faces can be used for surveillance and forensics purposes
Use Jarvis Smart cameras to See in Smarter Ways
Compared with traditional video motion detection, Jarvis’s intelligent VCA enables the camera to collect more valuable scene detection data based on frame differences.
Running as an edge-based application, Smart VCA also helps reduce bandwidth and storage usage.
If it is necessary to seek evidence,