Motion Detection Activation Before Inference

Motion Detection Activation Before Inference

Hello NVIDIA Developers,

I am working on a DeepStream pipeline and need to trigger inference based on motion detection. The idea is to minimize processing by running inference only when motion is detected in the video frame.

Requirements:

  • Motion detection should act as a pre-processing step.
  • nvinfer should only perform inference when motion is detected.
  • If motion detection can provide bounding box coordinates for regions, that would be helpful.

I would appreciate sample code, configuration guidance, or pointers to relevant plugins to achieve this.

Please customized with the Gst-nvdspreprocess (Alpha) — DeepStream documentation

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.