ocv_overlay

“ocv_overlay”* filter is used to draw the inference results on top of the original video, such as the detection bounding box, classified names, numbers or other labels.

Filter ocv_overlay
  Draw rectangle and text using libopencv.
    Inputs:
       #0: default (video)
    Outputs:
       #0: default (video)
ocv_overlay AVOptions:
  content_mode      <string>     ..FV..... basic or detailed content display (default "detailed")

*note: filter “ocv_overlay” requires OpenCV installed and add option --enable-libopencv when run the “configure” script before compiling the FFmpeg source code.