Documentation for Axxon One 2.0. Documentation for other versions of Axxon One is available too.

Previous page Human pose detector  Video stream and scene requirements for the Human pose detector and its sub-detectors Next page

DescriptionHuman pose detector uses a neural network and sub-detectors to determine each person's skeleton and detect poses that can indicate a potentially dangerous situation
Sub-detectors

You can create the following sub-detectors on the basis of the Human pose detector:

  1. Sitting person detection—detector generates an event when there is a sitting person in the frame.
  2. Person down detection—detector generates an event when there is a prostrate person in the frame.
  3. Hands up detection—detector generates an event when there is a person with one or two hands raised in the frame. A hand is considered raised if the forearm is parallel to the back.
  4. Handrail holding detection—detector generates an event when there is a person in the specified part of the frame who doesn't hold any of the specified handrails.
  5. People counter—detector counts objects in the specified area. Detector generates an event when the number of objects reaches or exceeds the specified limit.
  6. Close-standing people detection—detector generates an event when the distance between people in the frame exceeds the specified minimum value.
  7. People masking—detector masks people in the video image, it doesn't generate an event.

    Attention!

    Starting with DetectorPack 3.13, the People masking is no longer supported. We replaced it with the Track masking.

  8. Standard sub-detectors
Add-onsFor the Human pose detector and its sub-detectors to work, you must install Neuro Pack add-on (see Installing DetectorPack add-ons)
  • No labels