AI Face Detection & Recognition

FACE DETECTION Vs FACE RECOGNITION

Face Detection and Face Recognition are two different AI functions that capture human faces for use in surveillance applications. Face Detection uses AI machine learning to detect faces in a scene, enabling search by face for rapid footage review. Face Recognition takes this detected face and compares it to a database of stored faces on an AI video recorder. This system can then trigger alerts for known or unknown faces, perfect for boosting security and controlling access.

FEATURE OVERVIEW

  • Face Detection and Face Search: The camera or video recorder analyses the incoming video stream to detect faces. If detected, the system can capture a snapshot, stamp the playback footage or trigger an alarm. You can also search by detected faces, sorted by time of day.
  • Face Recognition: When pairing a camera with an AI video recorder, you can compare detected faces to a database of previously captured or imported faces. Accuracy thresholds can be set to prevent false alarms, eg: must be over 90% match tolerance.
  • Blocklist or Stranger Detection: In a blocklist configuration, if a detected face does match the face database, the system can prevent access or trigger an alarm. Similarly, in a stranger detection configuration, the system checks if a detected face does not match those in your database. This can prevent unknown people from accessing areas they are not allowed to enter, eg: staff-only areas, high security areas, etc.

IMPLEMENTATION

  • Activate Face Detection for desired cameras via the recorder or camera interface
  • Select or import faces to Face Recognition database from previously captured Face Detection events or from an imported image file via USB drive
  • Configure face database for Blocklist or Stranger Detection applications; multiple face databases can be made for different cameras and applications
  • Create triggers for detected and/or recognised faces, eg: take snapshot, trigger alarm, etc.

TECHNICAL CONSIDERATIONS

  • Camera & Recorder Performance: When performed solely by a video recorder, the Face Recognition process is computationally intense. The recorder first needs to analyse the video stream, which could be high resolution & high frame rate, to detect faces. It then needs to mark up face attributes for database comparison. Finally, it needs to search the database to find any accurate matches.
  • To improve performance, you can use Face Detection cameras to handle the video stream analysis and face attribute mark up, leaving the recorder to handle database comparisons. By using Face Detection cameras in tandem with a Face Recognition capable AI recorder, you can increase the number of cameras performing Face Recognition.

VIP Ai range of products.

 

DETECT & RECOGNISE FRIEND OR FOE & ALSO VEHICLE NUMBER PLATE ID

  • Machine learning people & vehicle detection: The camera or NVR is trained on hours of footage, to know what a person or vehicle looks like, enabling you to receive people & vehicle detection alerts to your phone, tablet or PC
  • Ignore what’s not important: Using machine learning, this AI feature prevents unwanted CCTV detection notifications, eg: trees in the wind, changing light, small animals, etc.
  • Don’t miss what’s important: Meaningful alerts help you action events that you may otherwise disregard as a false alarm or an unimportant notification.
  • Save yourself hours when searching footage: Continuous 24/7 CCTV footage stored on your NVR is stamped with these people & vehicle detection events, enabling you to find the footage you want faster.

Ai FEATURES MACHINE LEARNING FOR IMPROVING RESULTS

AI lets your CCTV system go beyond surveillance, delivering advanced & powerful detection tools. When properly configured, these functions can self analyse and alert you to any suspicious activity in real time via email/smartphone notification. Ai also assists you to easily search through surveillance footage to review event history.
AI People & Vehicle Detection is one of the most useful CCTV surveillance artificial intelligence technologies offered in the Watchguard range. Using machine learning, the camera or video recorder can distinguish between human and vehicle shapes to the exclusion of all else. This technology truly enhances the usefulness of CCTV for security and to dramatically save time in pinpointing video of interest quickly and with ease of convenience.

Ai GREATLY IMPROVES ACCESS TO EVENTS WITH EASY AND CONVENIENCE

Ai SMART FUNCTIONS:

  • Smart Motion Detection – is an AI filter (provided via Perimeter Protection) for Motion Detection that will only trigger an event if a human/vehicle is detected.
  • Tripwire – A virtual line is drawn over the camera image. If it gets broken, an event will be triggered. Suitable for both indoor and outdoor locations, with the camera mounted up high, looking down. AI functionality can be added to the tripwire via Perimeter Protection to enable events to trigger for people and/or vehicles only.
  • Intrusion – A virtual area is drawn over the image. If it gets broken, an event will be triggered. Suitable for both indoor and outdoor locations, with the camera mounted up high, looking down. AI functionality can be added to the intrusion via Perimeter Protection to enable events to trigger for people and/or vehicles only.
  • Fast-Moving – A virtual area is drawn over the image. If a fast-moving object is detected, an event will be triggered. Best suited for outdoor environments such as driveways.
  • Parking Detection – A virtual area is drawn over the image. If a vehcle parking is detected, an event will be triggered.
  • Crowd Gathering – A virtual area is drawn over the image. If a group of people is detected entering the area, an event will be triggered.
  • Loitering - A virtual area is drawn over the image. If someone is standing in the area for a set amount of time an event will be triggered. Best suited to a shop.

Smart Object Detection:

  • Abandoned – A virtual area is drawn over the image. If an object gets left in the area and event will be triggered.
  • Missing – A virtual area is drawn over the image. If an object is missing in the area, an event will be triggered.
  • Face Detection – Face Detection will trigger whenever any face has been detected. Best suited for use directly above entry doors.
  • Face Recognition – Trigger events upon detecting specific pre-determined faces from a face database. Face Detection must be used in conjunction with Face Recognition.
  • People Counting – Counts the number of people who cross a line (Line Crossing), are currently in an area (Region) or currently in an area and how long they’ve been there for (Queuing).
  • ANPR – Automatic License Plate Recognition – Used to trigger events after a number plate is detected.
  • Heat Map – A colour gradient overlay indicating the level of human traffic in each area of the image over a set period of time.

 

Ai Complete Range

Ai VIP New Arrivals

Deterrence