How it works
The Predator detection node is an application which can receive static images, pre-recorded video or live video streams and perform object recognition on the images using pre-trained neural networks.
Detected objects are highlighted in the image using a bounding box displaying the detected label (what the model thought the object was) and the confidence percentage (how confident the model was that the object it detected was the object it thought it was.)
Static images of the detection are saved for later use. Videos can also be recorded for playback displaying the object detection. Users can define what objects they wish to detect and save images of intruders. These objects are then highlighted in red.