Visual Cortek’s Virtual Guard is a software suite that includes features that are based on the most recent advances in the field of visual surveillance and event recognition.

Product Features

The Visual Cortek Virtual Guard memorizes the identity of a person and tracks them over time.

Visual Cortek’s Face Logger can assess in real-time the quality of captured faces and build a database of people's identities that is both concise and complete.

Using face detection and analysis technologies, good quality portraits are automatically captured allowing clear identification of the persons involved in occurring activities.

The system can be coupled to a face recognition module in order to obtain pro-active recognition of people in the scene

A sample of face images detected in video sequences.
Images are arranged from lowest quality to highest quality, when read from left to right.

The system offers scalable multi-camera coordination where fixed cameras control Pan-Tilt-Zoom cameras in order to capture the relevant visual information at the proper resolution.

The unique flexible and scalable Visual Cortek architecture allows collaboration between cameras in many forms. For instance, one fixed camera can control two PTZ cameras or more, such that when multiple events are occurring, each camera can be assigned to follow one of these simultaneous events. Reciprocally, large scene coverage can be ensured through the use of several fixed cameras controlling one or few PTZ cameras that would then alternate between multi-events.

This unique feature is made possible by the introduction of the Visual Cortek’s Instantaneous Moving Foreground Extraction method that allows quick motion tracking where PTZ cameras can zoom and follow objects in displacement.

The Visual Cortek Virtual Guard controls your cameras to ensure both close-up capture of occurring activities and global overview of the scene to make sure no important events are missed.

This learn - by - example feature of Visual Cortek’s system is the first application of the most recent advances in object recognition and machine learning in the field of visual surveillance and event recognition.

The Visual Cortek’s Virtual Guard alerts its users and receives instructions from them.

• Detected events of interest are reported directly to the remote users.
• Based on the user feedback, the system will adapt and automatically classify the events such that future events are reported in accordance to the user policies.

Visual Cortek uses selective detection functionalities developed at the VIVA lab to guarantee accurate detection of the elements of interest in busy scenes while ignoring other types of activities.  

The Visual Cortek Virtual Guard records and annotates the events with appropriate information.

• The system collects in real-time multiple descriptors about the moving entities’ physical attributes (size, shape, color signature, etc) and actions (speed, direction, position, etc) in order to characterize and categorize them.
• Using the extracted information, events are automatically classified and tagged with information concerning the video content.