Object Detection and Tracking
Scylla AI-powered proprietary object detection and classification engine augments your security infrastructure and provides situational awareness in real-time.

How it works


Scylla proprietary Object detection algorithm is trained to detect a number of specified objects in a single or multiple frames. The initial detection is fast and it serves to raise suspicion and forward the region of interest for more accurate assessment to the Charon classifier.

An alert is then triggered and distributed to the end user through Scylla web, mobile or integrated VMS channels along with the time, location and screenshot of the detection.

Scylla object detection is unaffected by the background scene or motion and can accurately analyze the content of video streams from stationary and moving cameras.

Scylla Object Tracking monitors object movement across a series of video frames.

Tracking of the same object across consequent frames enables multiple assessment and further boosts the accuracy of Scylla AI physical security solutions for real-time video analytics.

Scylla AI detects
The system is trained to detect and identify a wide variety of suspicious objects, weapons and personal protective equipment.
Weapon detection

Rifle

Shotgun

Gun

Knife & More
Personal protective equipment monitoring

Helmet

Mask & More
What makes Scylla Object Detection & Tracking System stand out





AI methodologies work autonomously 24/7 and self-improve
Can be seamlessly integrated with the majority of cameras and VMS
Effectively operates on cameras with moving backgrounds such as drones and body cams
Utilizes integrated zoom in & tracking algorithm that enables detection of distant objects
7 times lower hardware expectations compared to similar solutions on the market






Can be deployed both on-premise or on cloud
The object tracking module is based on a proprietary algorithm that is lightweight, accurate, and versatile.
The AI model is custom developed to re-identify specific types of objects with unparalleled accuracy.
The algorithm can be seamlessly applied to centralized solutions that analyze hundreds of video streams simultaneously.